1 systemd System and Service Manager
5 * As an experimental feature, udev now tries to lock the
6 disk device node (flock(LOCK_SH|LOCK_NB)) while it
7 executes events for the disk or any of its partitions.
8 Applications like partitioning programs can lock the
9 disk device node (flock(LOCK_EX)) and claim temporary
10 device ownership that way; udev will entirely skip all event
11 handling for this disk and its partitions. If the disk
12 was opened for writing, the close will trigger a partition
13 table rescan in udev's "watch" facility, and if needed
14 synthesize "change" events for the disk and all its partitions.
15 This is now unconditionally enabled, if it turns out to
16 cause major problems, we might turn it on only for specific
17 devices, or might need to disable it entirely. Device-mapper
18 devices are excluded from this logic.
20 * We temporarily dropped the "-l" switch for fsck invocations,
21 since they collide with the flock() logic above. util-linux
22 upstream has been changed already to avoid this conflict,
23 and we will readd "-l" as soon as util-linux with this
24 change has been released.
26 * The dependency on libattr has been removed. Since a long
27 time the extended attribute calls have moved to glibc, and
28 libattr is thus unnecessary.
30 * Virtualization detection works without priviliges now. This
31 means the systemd-detect-virt binary no longer requires
32 CAP_SYS_PTRACE file capabilities, and our daemons can run
33 with fewer privileges.
35 * systemd-networkd now runs under its own "systemd-network"
36 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
37 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
38 loses the ability to write to files owned by root this way.
40 * Similar, systemd-resolved now runs under its own
41 "systemd-resolve" user with no capabilities remaining.
43 * Similar, systemd-bus-proxyd now runs under its own
44 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
46 * systemd-networkd gained support for setting up "veth"
47 virtual ethernet devices for container connectivity, as well
48 as GRE and VTI tunnels.
50 * systemd-networkd will no longer automatically attempt to
51 manually load kernel modules necessary for certain tunnel
52 transports. Instead it is assumed the kernel loads them
53 automatically when required. This only works correctly on
54 very new kernels. On older kernels, please consider adding
55 the kernel modules to /etc/load-modules.d/ as a work-around.
57 * Two new service settings ProtectedHome= and ProtectedSystem=
58 have been added. When enabled they will make the user data
59 (such as /home) inaccessible or read-only and the system
60 (such as /usr) read-only, for specific services. This allows
61 very light-weight per-service sandboxing to avoid
62 modifications of user data or system files from
63 services. These two new switches have been enabled for all
64 of systemd's long-running services, where appropriate.
66 * Socket units gained new SocketUser= and SocketGroup=
67 settings to set the owner user and group of AF_UNIX sockets
68 and FIFOs in the file system.
70 * Socket units gained a new RemoveOnStop= setting. If enabled
71 all FIFOS and sockets in the file system will be removed
72 when the specific socket unit is stopped.
74 * Socket units gained a new Symlinks= setting. It takes a list
75 of symlinks to create to file system sockets or FIFOs
76 created by the specific unix sockets. This is useful to
77 manage symlinks to socket nodes with the same life-cycle as
80 * The /dev/log socket and /dev/initctl FIFO have been moved to
81 /run, and have been replaced by symlinks. This allows
82 connecting to these facilities even if PrivateDevices=yes is
83 used for a service (which makes /dev/log itself unavailable,
84 but /run is left). This also has the benefit of ensuring
85 that /dev only contains device nodes, directories and
86 symlinks, and nothing else.
88 * sd-daemon gained two new calls sd_pid_notify() and
89 sd_pid_notifyf(). They are similar to sd_notify() and
90 sd_notifyf(), but allow overriding of the source PID of
91 notification messages if permissions permit this. This is
92 useful to send notify messages on behalf of a different
93 process (for example, the parent process). The
94 systemd-notify tool has been updated to make use of this
95 when sending messages (so that notification messages now
96 originate from the shell script invoking systemd-notify and
97 not the systemd-notify process itself. This should minimize
98 a race where systemd fails to associate notification
99 messages to services when the originating process already
102 * A new "on-abnormal" setting for Restart= has been added. If
103 set it will result in automatic restarts on all "abnormal"
104 reasons for a process to exit, which includes unclean
105 signals, core dumps, timeouts and watchdog timeouts, but
106 does not include clean and unclean exit codes or clean
107 signals. Restart=on-abnormal is an alternative for
108 Restart=on-failure for services that shall be able to
109 terminate and avoid restarts on certain errors, by
110 indicating so with an unclean exit code. Restart=on-failure
111 or Restart=on-abnormal is now the recommended setting for
112 all long-running services.
114 * If the InaccessibleDirectories= service setting points to a
115 mount point (or if there are any submounts contained within
116 it), it is now attempted to completely unmount it, to make
117 the file systems truly unavailable for the respective
120 * The ReadOnlyDirectories= service setting and
121 systemd-nspawn's --read-only parameter are now recursively
122 applied to all submounts, too.
124 * Mount units may now be created transiently via the bus APIs.
126 * The support for SysV and LSB init scripts has been removed
127 from the systemd daemon itself. Instead, it is now
128 implemented as a generator that creates native systemd units
129 from these scripts when needed. This enables us to remove a
130 substantial amount of legacy code from PID 1, following the
131 fact that many distributions only ship a very small number
132 of LSB/SysV init scripts nowadays.
134 * Priviliged Xen (dom0) domains are not considered
135 virtualization anymore by the virtualization detection
136 logic. After all, they generally have unrestricted access to
137 the hardware and usually are used to manage the unprivileged
140 * systemd-tmpfiles gained a new "C" line type, for copying
141 files or entire directories.
143 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
144 lines. So far they have been non-globbing versions of the
145 latter, and have thus been redundant. In future it is
146 recommended to only use "z"; and "m" has hence been removed
147 from the documentation, even though it stays supported.
149 * A tmpfiles snippet to recreate the most basic structure in
150 /var has been added. This is enough to create the /var/run →
151 /run symlink and create a couple of structural
152 directories. This allows systems to boot up with an empty or
153 volatile /var. Of course, while with this change the core OS
154 now is capable with dealing with a volatile /var not all
155 user services are ready for it. However, we hope that sooner
156 or later many service daemons will be changed upstream so
157 that they are able to automatically create their necessary
158 directories in /var at boot, should they be missing. This is
159 the first step to allow state-less systems that only require
160 the vendor image for /usr to boot.
162 * systemd-nspawn has gained a new --tmpfs= switch to mount an
163 empty tmpfs instance to a specific directory. This is
164 particularly useful for making use of the automatic
165 reconstruction of /var (see above), by passing --tmpfs=/var.
167 * Access modes specified in tmpfiles snippets may now be
168 prefixed with "~", which indicates that they shall be masked
169 by whether the existing file or directly is currently
170 writable, readable or executable at all. Also, if specified
171 the sgid/suid/sticky bits will be masked for all
174 * A new passive target unit "network-pre.target" has been
175 added which is useful for services that shall run before any
176 network is configured, for example firewall scripts.
178 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
179 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
180 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
181 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
182 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
183 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
186 -- Berlin, 2014-06-11
190 * A new "systemd-timesyncd" daemon has been added for
191 synchronizing the system clock across the network. It
192 implements an SNTP client. In contrast to NTP
193 implementations such as chrony or the NTP reference server
194 this only implements a client side, and does not bother with
195 the full NTP complexity, focusing only on querying time from
196 one remote server and synchronizing the local clock to
197 it. Unless you intend to serve NTP to networked clients or
198 want to connect to local hardware clocks this simple NTP
199 client should be more than appropriate for most
200 installations. The daemon runs with minimal privileges, and
201 has been hooked up with networkd to only operate when
202 network connectivity is available. The daemon saves the
203 current clock to disk every time a new NTP sync has been
204 acquired, and uses this to possibly correct the system clock
205 early at bootup, in order to accommodate for systems that
206 lack an RTC such as the Raspberry Pi and embedded devices,
207 and make sure that time monotonically progresses on these
208 systems, even if it is not always correct. To make use of
209 this daemon a new system user and group "systemd-timesync"
210 needs to be created on installation of systemd.
212 * The queue "seqnum" interface of libudev has been disabled, as
213 it was generally incompatible with device namespacing as
214 sequence numbers of devices go "missing" if the devices are
215 part of a different namespace.
217 * "systemctl list-timers" and "systemctl list-sockets" gained
218 a --recursive switch for showing units of these types also
219 for all local containers, similar in style to the already
220 supported --recursive switch for "systemctl list-units".
222 * A new RebootArgument= setting has been added for service
223 units, which may be used to specify a kernel reboot argument
224 to use when triggering reboots with StartLimitAction=.
226 * A new FailureAction= setting has been added for service
227 units which may be used to specify an operation to trigger
228 when a service fails. This works similarly to
229 StartLimitAction=, but unlike it controls what is done
230 immediately rather than only after several attempts to
231 restart the service in question.
233 * hostnamed got updated to also expose the kernel name,
234 release, and version on the bus. This is useful for
235 executing commands like hostnamectl with the -H switch.
236 systemd-analyze makes use of this to properly display
237 details when running non-locally.
239 * The bootchart tool can now show cgroup information in the
242 * The CFS CPU quota cgroup attribute is now exposed for
243 services. The new CPUQuota= switch has been added for this
244 which takes a percentage value. Setting this will have the
245 result that a service may never get more CPU time than the
246 specified percentage, even if the machine is otherwise idle.
248 * systemd-networkd learned IPIP and SIT tunnel support.
250 * LSB init scripts exposing a dependency on $network will now
251 get a dependency on network-online.target rather than simply
252 network.target. This should bring LSB handling closer to
253 what it was on SysV systems.
255 * A new fsck.repair= kernel option has been added to control
256 how fsck shall deal with unclean file systems at boot.
258 * The (.ini) configuration file parser will now silently
259 ignore sections whose name begins with "X-". This may be
260 used to maintain application-specific extension sections in unit
263 * machined gained a new API to query the IP addresses of
264 registered containers. "machinectl status" has been updated
265 to show these addresses in its output.
267 * A new call sd_uid_get_display() has been added to the
268 sd-login APIs for querying the "primary" session of a
269 user. The "primary" session of the user is elected from the
270 user's sessions and generally a graphical session is
271 preferred over a text one.
273 * A minimal systemd-resolved daemon has been added. It
274 currently simply acts as a companion to systemd-networkd and
275 manages resolv.conf based on per-interface DNS
276 configuration, possibly supplied via DHCP. In the long run
277 we hope to extend this into a local DNSSEC enabled DNS and
280 * The systemd-networkd-wait-online tool is now enabled by
281 default. It will delay network-online.target until a network
282 connection has been configured. The tool primarily integrates
283 with networkd, but will also make a best effort to make sense
284 of network configuration performed in some other way.
286 * Two new service options StartupCPUShares= and
287 StartupBlockIOWeight= have been added that work similarly to
288 CPUShares= and BlockIOWeight= however only apply during
289 system startup. This is useful to prioritize certain services
290 differently during bootup than during normal runtime.
292 * hostnamed has been changed to prefer the statically
293 configured hostname in /etc/hostname (unless set to
294 'localhost' or empty) over any dynamic one supplied by
295 dhcp. With this change the rules for picking the hostname
296 match more closely the rules of other configuration settings
297 where the local administrator's configuration in /etc always
298 overrides any other settings.
300 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
301 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
302 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
303 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
304 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
305 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
306 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
307 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
308 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
309 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
310 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
311 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
312 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
313 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
314 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
315 Lindskog, WaLyong Cho, Will Woods, Zbigniew
318 -- Beijing, 2014-05-28
322 * When restoring the screen brightness at boot, stay away from
323 the darkest setting or from the lowest 5% of the available
324 range, depending on which is the larger value of both. This
325 should effectively protect the user from rebooting into a
326 black screen, should the brightness have been set to minimum
329 * sd-login gained a new sd_machine_get_class() call to
330 determine the class ("vm" or "container") of a machine
331 registered with machined.
333 * sd-login gained new calls
334 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
335 to query the identity of the peer of a local AF_UNIX
336 connection. They operate similarly to their sd_pid_get_xyz()
339 * PID 1 will now maintain a system-wide system state engine
340 with the states "starting", "running", "degraded",
341 "maintenance", "stopping". These states are bound to system
342 startup, normal runtime, runtime with at least one failed
343 service, rescue/emergency mode and system shutdown. This
344 state is shown in the "systemctl status" output when no unit
345 name is passed. It is useful to determine system state, in
346 particularly when doing so for many systems or containers at
349 * A new command "list-machines" has been added to "systemctl"
350 that lists all local OS containers and shows their system
351 state (see above), if systemd runs inside of them.
353 * systemctl gained a new "-r" switch to recursively enumerate
354 units on all local containers, when used with the
355 "list-unit" command (which is the default one that is
356 executed when no parameters are specified).
358 * The GPT automatic partition discovery logic will now honour
359 two GPT partition flags: one may be set on a partition to
360 cause it to be mounted read-only, and the other may be set
361 on a partition to ignore it during automatic discovery.
363 * Two new GPT type UUIDs have been added for automatic root
364 partition discovery, for 32-bit and 64-bit ARM. This is not
365 particularly useful for discovering the root directory on
366 these architectures during bare-metal boots (since UEFI is
367 not common there), but still very useful to allow booting of
368 ARM disk images in nspawn with the -i option.
370 * MAC addresses of interfaces created with nspawn's
371 --network-interface= switch will now be generated from the
372 machine name, and thus be stable between multiple invocations
375 * logind will now automatically remove all IPC objects owned
376 by a user if she or he fully logs out. This makes sure that
377 users who are logged out cannot continue to consume IPC
378 resources. This covers SysV memory, semaphores and message
379 queues as well as POSIX shared memory and message
380 queues. Traditionally, SysV and POSIX IPC had no life-cycle
381 limits. With this functionality, that is corrected. This may
382 be turned off by using the RemoveIPC= switch of logind.conf.
384 * The systemd-machine-id-setup and tmpfiles tools gained a
385 --root= switch to operate on a specific root directory,
388 * journald can now forward logged messages to the TTYs of all
389 logged in users ("wall"). This is the default for all
390 emergency messages now.
392 * A new tool systemd-journal-remote has been added to stream
393 journal log messages across the network.
395 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
396 controller trees are mounted into it. Note that the
397 directories mounted beneath it are not read-only. This is a
398 security measure and is particularly useful because glibc
399 actually includes a search logic to pick any tmpfs it can
400 find to implement shm_open() if /dev/shm is not available
401 (which it might very well be in namespaced setups).
403 * machinectl gained a new "poweroff" command to cleanly power
404 down a local OS container.
406 * The PrivateDevices= unit file setting will now also drop the
407 CAP_MKNOD capability from the capability bound set, and
408 imply DevicePolicy=closed.
410 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
411 comprehensively on all long-running systemd services where
414 * systemd-udevd will now run in a disassociated mount
415 namespace. To mount directories from udev rules, make sure to
416 pull in mount units via SYSTEMD_WANTS properties.
418 * The kdbus support gained support for uploading policy into
419 the kernel. sd-bus gained support for creating "monitoring"
420 connections that can eavesdrop into all bus communication
421 for debugging purposes.
423 * Timestamps may now be specified in seconds since the UNIX
424 epoch Jan 1st, 1970 by specifying "@" followed by the value
427 * Native tcpwrap support in systemd has been removed. tcpwrap
428 is old code, not really maintained anymore and has serious
429 shortcomings, and better options such as firewalls
430 exist. For setups that require tcpwrap usage, please
431 consider invoking your socket-activated service via tcpd,
432 like on traditional inetd.
434 * A new system.conf configuration option
435 DefaultTimerAccuracySec= has been added that controls the
436 default AccuracySec= setting of .timer units.
438 * Timer units gained a new WakeSystem= switch. If enabled,
439 timers configured this way will cause the system to resume
440 from system suspend (if the system supports that, which most
443 * Timer units gained a new Persistent= switch. If enabled,
444 timers configured this way will save to disk when they have
445 been last triggered. This information is then used on next
446 reboot to possible execute overdue timer events, that
447 could not take place because the system was powered off.
448 This enables simple anacron-like behaviour for timer units.
450 * systemctl's "list-timers" will now also list the time a
451 timer unit was last triggered in addition to the next time
452 it will be triggered.
454 * systemd-networkd will now assign predictable IPv4LL
455 addresses to its local interfaces.
457 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
458 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
459 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
460 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
461 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
462 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
463 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
464 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
467 -- Berlin, 2014-03-25
471 * A new unit file setting RestrictAddressFamilies= has been
472 added to restrict which socket address families unit
473 processes gain access to. This takes address family names
474 like "AF_INET" or "AF_UNIX", and is useful to minimize the
475 attack surface of services via exotic protocol stacks. This
476 is built on seccomp system call filters.
478 * Two new unit file settings RuntimeDirectory= and
479 RuntimeDirectoryMode= have been added that may be used to
480 manage a per-daemon runtime directories below /run. This is
481 an alternative for setting up directory permissions with
482 tmpfiles snippets, and has the advantage that the runtime
483 directory's lifetime is bound to the daemon runtime and that
484 the daemon starts up with an empty directory each time. This
485 is particularly useful when writing services that drop
486 privileges using the User= or Group= setting.
488 * The DeviceAllow= unit setting now supports globbing for
489 matching against device group names.
491 * The systemd configuration file system.conf gained new
492 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
493 DefaultMemoryAccounting= to globally turn on/off accounting
494 for specific resources (cgroups) for all units. These
495 settings may still be overridden individually in each unit
498 * systemd-gpt-auto-generator is now able to discover /srv and
499 root partitions in addition to /home and swap partitions. It
500 also supports LUKS-encrypted partitions now. With this in
501 place, automatic discovery of partitions to mount following
502 the Discoverable Partitions Specification
503 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
504 is now a lot more complete. This allows booting without
505 /etc/fstab and without root= on the kernel command line on
506 systems prepared appropriately.
508 * systemd-nspawn gained a new --image= switch which allows
509 booting up disk images and Linux installations on any block
510 device that follow the Discoverable Partitions Specification
511 (see above). This means that installations made with
512 appropriately updated installers may now be started and
513 deployed using container managers, completely
514 unmodified. (We hope that libvirt-lxc will add support for
515 this feature soon, too.)
517 * systemd-nspawn gained a new --network-macvlan= setting to
518 set up a private macvlan interface for the
519 container. Similarly, systemd-networkd gained a new
520 Kind=macvlan setting in .netdev files.
522 * systemd-networkd now supports configuring local addresses
525 * A new tool systemd-network-wait-online has been added to
526 synchronously wait for network connectivity using
529 * The sd-bus.h bus API gained a new sd_bus_track object for
530 tracking the life-cycle of bus peers. Note that sd-bus.h is
531 still not a public API though (unless you specify
532 --enable-kdbus on the configure command line, which however
533 voids your warranty and you get no API stability guarantee).
535 * The $XDG_RUNTIME_DIR runtime directories for each user are
536 now individual tmpfs instances, which has the benefit of
537 introducing separate pools for each user, with individual
538 size limits, and thus making sure that unprivileged clients
539 can no longer negatively impact the system or other users by
540 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
541 RuntimeDirectorySize= has been introduced that allows
542 controlling the default size limit for all users. It
543 defaults to 10% of the available physical memory. This is no
544 replacement for quotas on tmpfs though (which the kernel
545 still does not support), as /dev/shm and /tmp are still
546 shared resources used by both the system and unprivileged
549 * logind will now automatically turn off automatic suspending
550 on laptop lid close when more than one display is
551 connected. This was previously expected to be implemented
552 individually in desktop environments (such as GNOME),
553 however has been added to logind now, in order to fix a
554 boot-time race where a desktop environment might not have
555 been started yet and thus not been able to take an inhibitor
556 lock at the time where logind already suspends the system
559 * logind will now wait at least 30s after each system
560 suspend/resume cycle, and 3min after system boot before
561 suspending the system due to a closed laptop lid. This
562 should give USB docking stations and similar enough time to
563 be probed and configured after system resume and boot in
564 order to then act as suspend blocker.
566 * systemd-run gained a new --property= setting which allows
567 initialization of resource control properties (and others)
568 for the created scope or service unit. Example: "systemd-run
569 --property=BlockIOWeight=10 updatedb" may be used to run
570 updatedb at a low block IO scheduling weight.
572 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
573 now also work in --scope mode.
575 * When systemd is compiled with kdbus support, basic support
576 for enforced policies is now in place. (Note that enabling
577 kdbus still voids your warranty and no API compatibility
580 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
581 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
582 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
583 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
584 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
585 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
586 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
587 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
588 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
589 Zbigniew Jędrzejewski-Szmek
591 -- Berlin, 2014-03-12
595 * systemd will now relabel /dev after loading the SMACK policy
596 according to SMACK rules.
598 * A new unit file option AppArmorProfile= has been added to
599 set the AppArmor profile for the processes of a unit.
601 * A new condition check ConditionArchitecture= has been added
602 to conditionalize units based on the system architecture, as
603 reported by uname()'s "machine" field.
605 * systemd-networkd now supports matching on the system
606 virtualization, architecture, kernel command line, host name
609 * logind is now a lot more aggressive when suspending the
610 machine due to a closed laptop lid. Instead of acting only
611 on the lid close action, it will continuously watch the lid
612 status and act on it. This is useful for laptops where the
613 power button is on the outside of the chassis so that it can
614 be reached without opening the lid (such as the Lenovo
615 Yoga). On those machines, logind will now immediately
616 re-suspend the machine if the power button has been
617 accidentally pressed while the laptop was suspended and in a
620 * logind will now watch SW_DOCK switches and inhibit reaction
621 to the lid switch if it is pressed. This means that logind
622 will not suspend the machine anymore if the lid is closed
623 and the system is docked, if the laptop supports SW_DOCK
624 notifications via the input layer. Note that ACPI docking
625 stations do not generate this currently. Also note that this
626 logic is usually not fully sufficient and Desktop
627 Environments should take a lid switch inhibitor lock when an
628 external display is connected, as systemd will not watch
631 * nspawn will now make use of the devices cgroup controller by
632 default, and only permit creation of and access to the usual
633 API device nodes like /dev/null or /dev/random, as well as
634 access to (but not creation of) the pty devices.
636 * We will now ship a default .network file for
637 systemd-networkd that automatically configures DHCP for
638 network interfaces created by nspawn's --network-veth or
639 --network-bridge= switches.
641 * systemd will now understand the usual M, K, G, T suffixes
642 according to SI conventions (i.e. to the base 1000) when
643 referring to throughput and hardware metrics. It will stay
644 with IEC conventions (i.e. to the base 1024) for software
645 metrics, according to what is customary according to
646 Wikipedia. We explicitly document which base applies for
647 each configuration option.
649 * The DeviceAllow= setting in unit files now supports a syntax
650 to whitelist an entire group of devices node majors at once,
651 based on the /proc/devices listing. For example, with the
652 string "char-pts", it is now possible to whitelist all
653 current and future pseudo-TTYs at once.
655 * sd-event learned a new "post" event source. Event sources of
656 this type are triggered by the dispatching of any event
657 source of a type that is not "post". This is useful for
658 implementing clean-up and check event sources that are
659 triggered by other work being done in the program.
661 * systemd-networkd is no longer statically enabled, but uses
662 the usual [Install] sections so that it can be
663 enabled/disabled using systemctl. It still is enabled by
666 * When creating a veth interface pair with systemd-nspawn, the
667 host side will now be prefixed with "vb-" if
668 --network-bridge= is used, and with "ve-" if --network-veth
669 is used. This way, it is easy to distinguish these cases on
670 the host, for example to apply different configuration to
671 them with systemd-networkd.
673 * The compatibility libraries for libsystemd-journal.so,
674 libsystem-id128.so, libsystemd-login.so and
675 libsystemd-daemon.so do not make use of IFUNC
676 anymore. Instead, we now build libsystemd.so multiple times
677 under these alternative names. This means that the footprint
678 is drastically increased, but given that these are
679 transitional compatibility libraries, this should not matter
680 much. This change has been made necessary to support the ARM
681 platform for these compatibility libraries, as the ARM
682 toolchain is not really at the same level as the toolchain
683 for other architectures like x86 and does not support
684 IFUNC. Please make sure to use --enable-compat-libs only
685 during a transitional period!
687 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
688 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
689 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
690 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
691 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
692 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
693 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
694 Zbigniew Jędrzejewski-Szmek
696 -- Berlin, 2014-02-24
700 * A new component "systemd-networkd" has been added that can
701 be used to configure local network interfaces statically or
702 via DHCP. It is capable of bringing up bridges, VLANs, and
703 bonding. Currently, no hook-ups for interactive network
704 configuration are provided. Use this for your initrd,
705 container, embedded, or server setup if you need a simple,
706 yet powerful, network configuration solution. This
707 configuration subsystem is quite nifty, as it allows wildcard
708 hotplug matching in interfaces. For example, with a single
709 configuration snippet, you can configure that all Ethernet
710 interfaces showing up are automatically added to a bridge,
711 or similar. It supports link-sensing and more.
713 * A new tool "systemd-socket-proxyd" has been added which can
714 act as a bidirectional proxy for TCP sockets. This is
715 useful for adding socket activation support to services that
716 do not actually support socket activation, including virtual
717 machines and the like.
719 * Add a new tool to save/restore rfkill state on
722 * Save/restore state of keyboard backlights in addition to
723 display backlights on shutdown/boot.
725 * udev learned a new SECLABEL{} construct to label device
726 nodes with a specific security label when they appear. For
727 now, only SECLABEL{selinux} is supported, but the syntax is
728 prepared for additional security frameworks.
730 * udev gained a new scheme to configure link-level attributes
731 from files in /etc/systemd/network/*.link. These files can
732 match against MAC address, device path, driver name and type,
733 and will apply attributes like the naming policy, link speed,
734 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
735 address assignment policy (randomized, ...).
737 * The configuration of network interface naming rules for
738 "permanent interface names" has changed: a new NamePolicy=
739 setting in the [Link] section of .link files determines the
740 priority of possible naming schemes (onboard, slot, mac,
741 path). The default value of this setting is determined by
742 /usr/lib/net/links/99-default.link. Old
743 80-net-name-slot.rules udev configuration file has been
744 removed, so local configuration overriding this file should
745 be adapated to override 99-default.link instead.
747 * When the User= switch is used in a unit file, also
748 initialize $SHELL= based on the user database entry.
750 * systemd no longer depends on libdbus. All communication is
751 now done with sd-bus, systemd's low-level bus library
754 * kdbus support has been added to PID 1 itself. When kdbus is
755 enabled, this causes PID 1 to set up the system bus and
756 enable support for a new ".busname" unit type that
757 encapsulates bus name activation on kdbus. It works a little
758 bit like ".socket" units, except for bus names. A new
759 generator has been added that converts classic dbus1 service
760 activation files automatically into native systemd .busname
763 * sd-bus: add a light-weight vtable implementation that allows
764 defining objects on the bus with a simple static const
765 vtable array of its methods, signals and properties.
767 * systemd will not generate or install static dbus
768 introspection data anymore to /usr/share/dbus-1/interfaces,
769 as the precise format of these files is unclear, and
770 nothing makes use of it.
772 * A proxy daemon is now provided to proxy clients connecting
773 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
774 compatibility with classic D-Bus.
776 * A bus driver implementation has been added that supports the
777 classic D-Bus bus driver calls on kdbus, also for
778 compatibility purposes.
780 * A new API "sd-event.h" has been added that implements a
781 minimal event loop API built around epoll. It provides a
782 couple of features that direct epoll usage is lacking:
783 prioritization of events, scales to large numbers of timer
784 events, per-event timer slack (accuracy), system-wide
785 coalescing of timer events, exit handlers, watchdog
786 supervision support using systemd's sd_notify() API, child
789 * A new API "sd-rntl.h" has been added that provides an API
790 around the route netlink interface of the kernel, similar in
793 * A new API "sd-dhcp-client.h" has been added that provides a
794 small DHCPv4 client-side implementation. This is used by
797 * There is a new kernel command line option
798 "systemd.restore_state=0|1". When set to "0", none of the
799 systemd tools will restore saved runtime state to hardware
800 devices. More specifically, the rfkill and backlight states
803 * The FsckPassNo= compatibility option in mount/service units
804 has been removed. The fstab generator will now add the
805 necessary dependencies automatically, and does not require
806 PID1's support for that anymore.
808 * journalctl gained a new switch, --list-boots, that lists
809 recent boots with their times and boot IDs.
811 * The various tools like systemctl, loginctl, timedatectl,
812 busctl, systemd-run, ... have gained a new switch "-M" to
813 connect to a specific, local OS container (as direct
814 connection, without requiring SSH). This works on any
815 container that is registered with machined, such as those
816 created by libvirt-lxc or nspawn.
818 * systemd-run and systemd-analyze also gained support for "-H"
819 to connect to remote hosts via SSH. This is particularly
820 useful for systemd-run because it enables queuing of jobs
823 * machinectl gained a new command "login" to open a getty
824 login in any local container. This works with any container
825 that is registered with machined (such as those created by
826 libvirt-lxc or nspawn), and which runs systemd inside.
828 * machinectl gained a new "reboot" command that may be used to
829 trigger a reboot on a specific container that is registered
830 with machined. This works on any container that runs an init
833 * systemctl gained a new "list-timers" command to print a nice
834 listing of installed timer units with the times they elapse
837 * Alternative reboot() parameters may now be specified on the
838 "systemctl reboot" command line and are passed to the
839 reboot() system call.
841 * systemctl gained a new --job-mode= switch to configure the
842 mode to queue a job with. This is a more generic version of
843 --fail, --irreversible, and --ignore-dependencies, which are
844 still available but not advertised anymore.
846 * /etc/systemd/system.conf gained new settings to configure
847 various default timeouts of units, as well as the default
848 start limit interval and burst. These may still be overridden
851 * PID1 will now export on the bus profile data of the security
852 policy upload process (such as the SELinux policy upload to
855 * journald: when forwarding logs to the console, include
856 timestamps (following the setting in
857 /sys/module/printk/parameters/time).
859 * OnCalendar= in timer units now understands the special
860 strings "yearly" and "annually". (Both are equivalent)
862 * The accuracy of timer units is now configurable with the new
863 AccuracySec= setting. It defaults to 1min.
865 * A new dependency type JoinsNamespaceOf= has been added that
866 allows running two services within the same /tmp and network
867 namespace, if PrivateNetwork= or PrivateTmp= are used.
869 * A new command "cat" has been added to systemctl. It outputs
870 the original unit file of a unit, and concatenates the
871 contents of additional "drop-in" unit file snippets, so that
872 the full configuration is shown.
874 * systemctl now supports globbing on the various "list-xyz"
875 commands, like "list-units" or "list-sockets", as well as on
876 those commands which take multiple unit names.
878 * journalctl's --unit= switch gained support for globbing.
880 * All systemd daemons now make use of the watchdog logic so
881 that systemd automatically notices when they hang.
883 * If the $container_ttys environment variable is set,
884 getty-generator will automatically spawn a getty for each
885 listed tty. This is useful for container managers to request
886 login gettys to be spawned on as many ttys as needed.
888 * %h, %s, %U specifier support is not available anymore when
889 used in unit files for PID 1. This is because NSS calls are
890 not safe from PID 1. They stay available for --user
891 instances of systemd, and as special case for the root user.
893 * loginctl gained a new "--no-legend" switch to turn off output
896 * The "sd-login.h" API gained three new calls:
897 sd_session_is_remote(), sd_session_get_remote_user(),
898 sd_session_get_remote_host() to query information about
901 * The udev hardware database now also carries vendor/product
902 information of SDIO devices.
904 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
905 determine whether watchdog notifications are requested by
908 * Socket-activated per-connection services now include a
909 short description of the connection parameters in the
912 * tmpfiles gained a new "--boot" option. When this is not used,
913 only lines where the command character is not suffixed with
914 "!" are executed. When this option is specified, those
915 options are executed too. This partitions tmpfiles
916 directives into those that can be safely executed at any
917 time, and those which should be run only at boot (for
918 example, a line that creates /run/nologin).
920 * A new API "sd-resolve.h" has been added which provides a simple
921 asynchronous wrapper around glibc NSS host name resolution
922 calls, such as getaddrinfo(). In contrast to glibc's
923 getaddrinfo_a(), it does not use signals. In contrast to most
924 other asynchronous name resolution libraries, this one does
925 not reimplement DNS, but reuses NSS, so that alternate
926 host name resolution systems continue to work, such as mDNS,
927 LDAP, etc. This API is based on libasyncns, but it has been
928 cleaned up for inclusion in systemd.
930 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
931 "sd-daemon.h" are no longer found in individual libraries
932 libsystemd-journal.so, libsystemd-login.so,
933 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
934 merged them into a single library, libsystemd.so, which
935 provides all symbols. The reason for this is cyclic
936 dependencies, as these libraries tend to use each other's
937 symbols. So far, we have managed to workaround that by linking
938 a copy of a good part of our code into each of these
939 libraries again and again, which, however, makes certain
940 things hard to do, like sharing static variables. Also, it
941 substantially increases footprint. With this change, there
942 is only one library for the basic APIs systemd
943 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
944 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
945 library as well, however are subject to the --enable-kdbus
946 switch (see below). Note that "sd-dhcp-client.h" is not part
947 of this library (this is because it only consumes, never
948 provides, services of/to other APIs). To make the transition
949 easy from the separate libraries to the unified one, we
950 provide the --enable-compat-libs compile-time switch which
951 will generate stub libraries that are compatible with the
952 old ones but redirect all calls to the new one.
954 * All of the kdbus logic and the new APIs "sd-bus.h",
955 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
956 and "sd-utf8.h" are compile-time optional via the
957 "--enable-kdbus" switch, and they are not compiled in by
958 default. To make use of kdbus, you have to explicitly enable
959 the switch. Note however, that neither the kernel nor the
960 userspace API for all of this is considered stable yet. We
961 want to maintain the freedom to still change the APIs for
962 now. By specifying this build-time switch, you acknowledge
963 that you are aware of the instability of the current
966 * Also, note that while kdbus is pretty much complete,
967 it lacks one thing: proper policy support. This means you
968 can build a fully working system with all features; however,
969 it will be highly insecure. Policy support will be added in
970 one of the next releases, at the same time that we will
971 declare the APIs stable.
973 * When the kernel command-line argument "kdbus" is specified,
974 systemd will automatically load the kdbus.ko kernel module. At
975 this stage of development, it is only useful for testing kdbus
976 and should not be used in production. Note: if "--enable-kdbus"
977 is specified, and the kdbus.ko kernel module is available, and
978 "kdbus" is added to the kernel command line, the entire system
979 runs with kdbus instead of dbus-daemon, with the above mentioned
980 problem of missing the system policy enforcement. Also a future
981 version of kdbus.ko or a newer systemd will not be compatible with
982 each other, and will unlikely be able to boot the machine if only
983 one of them is updated.
985 * systemctl gained a new "import-environment" command which
986 uploads the caller's environment (or parts thereof) into the
987 service manager so that it is inherited by services started
988 by the manager. This is useful to upload variables like
989 $DISPLAY into the user service manager.
991 * A new PrivateDevices= switch has been added to service units
992 which allows running a service with a namespaced /dev
993 directory that does not contain any device nodes for
994 physical devices. More specifically, it only includes devices
995 such as /dev/null, /dev/urandom, and /dev/zero which are API
998 * logind has been extended to support behaviour like VT
999 switching on seats that do not support a VT. This makes
1000 multi-session available on seats that are not the first seat
1001 (seat0), and on systems where kernel support for VTs has
1002 been disabled at compile-time.
1004 * If a process holds a delay lock for system sleep or shutdown
1005 and fails to release it in time, we will now log its
1006 identity. This makes it easier to identify processes that
1007 cause slow suspends or power-offs.
1009 * When parsing /etc/crypttab, support for a new key-slot=
1010 option as supported by Debian is added. It allows indicating
1011 which LUKS slot to use on disk, speeding up key loading.
1013 * The sd_journald_sendv() API call has been checked and
1014 officially declared to be async-signal-safe so that it may
1015 be invoked from signal handlers for logging purposes.
1017 * Boot-time status output is now enabled automatically after a
1018 short timeout if boot does not progress, in order to give
1019 the user an indication what she or he is waiting for.
1021 * The boot-time output has been improved to show how much time
1022 remains until jobs expire.
1024 * The KillMode= switch in service units gained a new possible
1025 value "mixed". If set, and the unit is shut down, then the
1026 initial SIGTERM signal is sent only to the main daemon
1027 process, while the following SIGKILL signal is sent to
1028 all remaining processes of the service.
1030 * When a scope unit is registered, a new property "Controller"
1031 may be set. If set to a valid bus name, systemd will send a
1032 RequestStop() signal to this name when it would like to shut
1033 down the scope. This may be used to hook manager logic into
1034 the shutdown logic of scope units. Also, scope units may now
1035 be put in a special "abandoned" state, in which case the
1036 manager process which created them takes no further
1037 responsibilities for it.
1039 * When reading unit files, systemd will now verify
1040 the access mode of these files, and warn about certain
1041 suspicious combinations. This has been added to make it
1042 easier to track down packaging bugs where unit files are
1043 marked executable or world-writable.
1045 * systemd-nspawn gained a new "--setenv=" switch to set
1046 container-wide environment variables. The similar option in
1047 systemd-activate was renamed from "--environment=" to
1048 "--setenv=" for consistency.
1050 * systemd-nspawn has been updated to create a new kdbus domain
1051 for each container that is invoked, thus allowing each
1052 container to have its own set of system and user buses,
1053 independent of the host.
1055 * systemd-nspawn gained a new --drop-capability= switch to run
1056 the container with less capabilities than the default. Both
1057 --drop-capability= and --capability= now take the special
1058 string "all" for dropping or keeping all capabilities.
1060 * systemd-nspawn gained new switches for executing containers
1061 with specific SELinux labels set.
1063 * systemd-nspawn gained a new --quiet switch to not generate
1064 any additional output but the container's own console
1067 * systemd-nspawn gained a new --share-system switch to run a
1068 container without PID namespacing enabled.
1070 * systemd-nspawn gained a new --register= switch to control
1071 whether the container is registered with systemd-machined or
1072 not. This is useful for containers that do not run full
1073 OS images, but only specific apps.
1075 * systemd-nspawn gained a new --keep-unit which may be used
1076 when invoked as the only program from a service unit, and
1077 results in registration of the unit service itself in
1078 systemd-machined, instead of a newly opened scope unit.
1080 * systemd-nspawn gained a new --network-interface= switch for
1081 moving arbitrary interfaces to the container. The new
1082 --network-veth switch creates a virtual Ethernet connection
1083 between host and container. The new --network-bridge=
1084 switch then allows assigning the host side of this virtual
1085 Ethernet connection to a bridge device.
1087 * systemd-nspawn gained a new --personality= switch for
1088 setting the kernel personality for the container. This is
1089 useful when running a 32-bit container on a 64-bit host. A
1090 similar option Personality= is now also available for service
1093 * logind will now also track a "Desktop" identifier for each
1094 session which encodes the desktop environment of it. This is
1095 useful for desktop environments that want to identify
1096 multiple running sessions of itself easily.
1098 * A new SELinuxContext= setting for service units has been
1099 added that allows setting a specific SELinux execution
1100 context for a service.
1102 * Most systemd client tools will now honour $SYSTEMD_LESS for
1103 settings of the "less" pager. By default, these tools will
1104 override $LESS to allow certain operations to work, such as
1105 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1106 influence this logic.
1108 * systemd's "seccomp" hook-up has been changed to make use of
1109 the libseccomp library instead of using its own
1110 implementation. This has benefits for portability among
1113 * For usage together with SystemCallFilter=, a new
1114 SystemCallErrorNumber= setting has been introduced that
1115 allows configuration of a system error number to be returned
1116 on filtered system calls, instead of immediately killing the
1117 process. Also, SystemCallArchitectures= has been added to
1118 limit access to system calls of a particular architecture
1119 (in order to turn off support for unused secondary
1120 architectures). There is also a global
1121 SystemCallArchitectures= setting in system.conf now to turn
1122 off support for non-native system calls system-wide.
1124 * systemd requires a kernel with a working name_to_handle_at(),
1125 please see the kernel config requirements in the README file.
1127 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1128 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1129 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1130 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1131 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1132 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1133 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1134 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1135 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1136 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1137 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1138 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1139 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1140 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
1141 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
1142 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
1143 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
1144 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
1145 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
1146 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
1147 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
1148 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
1149 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
1150 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
1152 -- Berlin, 2014-02-20
1156 * logind has gained support for facilitating privileged input
1157 and drm device access for unprivileged clients. This work is
1158 useful to allow Wayland display servers (and similar
1159 programs, such as kmscon) to run under the user's ID and
1160 access input and drm devices which are normally
1161 protected. When this is used (and the kernel is new enough)
1162 logind will "mute" IO on the file descriptors passed to
1163 Wayland as long as it is in the background and "unmute" it
1164 if it returns into the foreground. This allows secure
1165 session switching without allowing background sessions to
1166 eavesdrop on input and display data. This also introduces
1167 session switching support if VT support is turned off in the
1168 kernel, and on seats that are not seat0.
1170 * A new kernel command line option luks.options= is understood
1171 now which allows specifiying LUKS options for usage for LUKS
1172 encrypted partitions specified with luks.uuid=.
1174 * tmpfiles.d(5) snippets may now use specifier expansion in
1175 path names. More specifically %m, %b, %H, %v, are now
1176 replaced by the local machine id, boot id, hostname, and
1177 kernel version number.
1179 * A new tmpfiles.d(5) command "m" has been introduced which
1180 may be used to change the owner/group/access mode of a file
1181 or directory if it exists, but do nothing if it does not.
1183 * This release removes high-level support for the
1184 MemorySoftLimit= cgroup setting. The underlying kernel
1185 cgroup attribute memory.soft_limit= is currently badly
1186 designed and likely to be removed from the kernel API in its
1187 current form, hence we should not expose it for now.
1189 * The memory.use_hierarchy cgroup attribute is now enabled for
1190 all cgroups systemd creates in the memory cgroup
1191 hierarchy. This option is likely to be come the built-in
1192 default in the kernel anyway, and the non-hierarchial mode
1193 never made much sense in the intrinsically hierarchial
1196 * A new field _SYSTEMD_SLICE= is logged along with all journal
1197 messages containing the slice a message was generated
1198 from. This is useful to allow easy per-customer filtering of
1199 logs among other things.
1201 * systemd-journald will no longer adjust the group of journal
1202 files it creates to the "systemd-journal" group. Instead we
1203 rely on the journal directory to be owned by the
1204 "systemd-journal" group, and its setgid bit set, so that the
1205 kernel file system layer will automatically enforce that
1206 journal files inherit this group assignment. The reason for
1207 this change is that we cannot allow NSS look-ups from
1208 journald which would be necessary to resolve
1209 "systemd-journal" to a numeric GID, because this might
1210 create deadlocks if NSS involves synchronous queries to
1211 other daemons (such as nscd, or sssd) which in turn are
1212 logging clients of journald and might block on it, which
1213 would then dead lock. A tmpfiles.d(5) snippet included in
1214 systemd will make sure the setgid bit and group are
1215 properly set on the journal directory if it exists on every
1216 boot. However, we recommend adjusting it manually after
1217 upgrades too (or from RPM scriptlets), so that the change is
1218 not delayed until next reboot.
1220 * Backlight and random seed files in /var/lib/ have moved into
1221 the /var/lib/systemd/ directory, in order to centralize all
1222 systemd generated files in one directory.
1224 * Boot time performance measurements (as displayed by
1225 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1226 performance information if that's available to determine how
1227 much time BIOS and boot loader initialization required. With
1228 a sufficiently new BIOS you hence no longer need to boot
1229 with Gummiboot to get access to such information.
1231 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1232 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1233 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1234 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1235 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1236 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1237 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1239 -- Berlin, 2013-10-02
1243 * The Restart= option for services now understands a new
1244 on-watchdog setting, which will restart the service
1245 automatically if the service stops sending out watchdog keep
1246 alive messages (as configured with WatchdogSec=).
1248 * The getty generator (which is responsible for bringing up a
1249 getty on configured serial consoles) will no longer only
1250 start a getty on the primary kernel console but on all
1251 others, too. This makes the order in which console= is
1252 specified on the kernel command line less important.
1254 * libsystemd-logind gained a new sd_session_get_vt() call to
1255 retrieve the VT number of a session.
1257 * If the option "tries=0" is set for an entry of /etc/crypttab
1258 its passphrase is queried indefinitely instead of any
1259 maximum number of tries.
1261 * If a service with a configure PID file terminates its PID
1262 file will now be removed automatically if it still exists
1263 afterwards. This should put an end to stale PID files.
1265 * systemd-run will now also take relative binary path names
1266 for execution and no longer insists on absolute paths.
1268 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1269 paths that are optionally prefixed with "-" to indicate that
1270 it should not be considered a failure if they do not exist.
1272 * journalctl -o (and similar commands) now understands a new
1273 output mode "short-precise", it is similar to "short" but
1274 shows timestamps with usec accuracy.
1276 * The option "discard" (as known from Debian) is now
1277 synonymous to "allow-discards" in /etc/crypttab. In fact,
1278 "discard" is preferred now (since it is easier to remember
1281 * Some licensing clean-ups were made, so that more code is now
1282 LGPL-2.1 licensed than before.
1284 * A minimal tool to save/restore the display backlight
1285 brightness across reboots has been added. It will store the
1286 backlight setting as late as possible at shutdown, and
1287 restore it as early as possible during reboot.
1289 * A logic to automatically discover and enable home and swap
1290 partitions on GPT disks has been added. With this in place
1291 /etc/fstab becomes optional for many setups as systemd can
1292 discover certain partitions located on the root disk
1293 automatically. Home partitions are recognized under their
1294 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1295 partitions are recognized under their GPT type ID
1296 0657fd6da4ab43c484e50933c84b4f4f.
1298 * systemd will no longer pass any environment from the kernel
1299 or initrd to system services. If you want to set an
1300 environment for all services, do so via the kernel command
1301 line systemd.setenv= assignment.
1303 * The systemd-sysctl tool no longer natively reads the file
1304 /etc/sysctl.conf. If desired, the file should be symlinked
1305 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1306 legacy support by a symlink rather than built-in code, it
1307 also makes the otherwise hidden order of application of the
1308 different files visible. (Note that this partly reverts to a
1309 pre-198 application order of sysctl knobs!)
1311 * The "systemctl set-log-level" and "systemctl dump" commands
1312 have been moved to systemd-analyze.
1314 * systemd-run learned the new --remain-after-exit switch,
1315 which causes the scope unit not to be cleaned up
1316 automatically after the process terminated.
1318 * tmpfiles learned a new --exclude-prefix= switch to exclude
1319 certain paths from operation.
1321 * journald will now automatically flush all messages to disk
1322 as soon as a message of the log priorities CRIT, ALERT or
1325 Contributions from: Andrew Cook, Brandon Philips, Christian
1326 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1327 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1328 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1329 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1330 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1331 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1332 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1333 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1334 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1335 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1336 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1337 William Giokas, Zbigniew Jędrzejewski-Szmek
1339 -- Berlin, 2013-09-13
1343 * The documentation has been updated to cover the various new
1344 concepts introduced with 205.
1346 * Unit files now understand the new %v specifier which
1347 resolves to the kernel version string as returned by "uname
1350 * systemctl now supports filtering the unit list output by
1351 load state, active state and sub state, using the new
1354 * "systemctl status" will now show the results of the
1355 condition checks (like ConditionPathExists= and similar) of
1356 the last start attempts of the unit. They are also logged to
1359 * "journalctl -b" may now be used to look for boot output of a
1360 specific boot. Try "journalctl -b -1" for the previous boot,
1361 but the syntax is substantially more powerful.
1363 * "journalctl --show-cursor" has been added which prints the
1364 cursor string the last shown log line. This may then be used
1365 with the new "journalctl --after-cursor=" switch to continue
1366 browsing logs from that point on.
1368 * "journalctl --force" may now be used to force regeneration
1371 * Creation of "dead" device nodes has been moved from udev
1372 into kmod and tmpfiles. Previously, udev would read the kmod
1373 databases to pre-generate dead device nodes based on meta
1374 information contained in kernel modules, so that these would
1375 be auto-loaded on access rather then at boot. As this
1376 does not really have much to do with the exposing actual
1377 kernel devices to userspace this has always been slightly
1378 alien in the udev codebase. Following the new scheme kmod
1379 will now generate a runtime snippet for tmpfiles from the
1380 module meta information and it now is tmpfiles' job to the
1381 create the nodes. This also allows overriding access and
1382 other parameters for the nodes using the usual tmpfiles
1383 facilities. As side effect this allows us to remove the
1384 CAP_SYS_MKNOD capability bit from udevd entirely.
1386 * logind's device ACLs may now be applied to these "dead"
1387 devices nodes too, thus finally allowing managed access to
1388 devices such as /dev/snd/sequencer whithout loading the
1389 backing module right-away.
1391 * A new RPM macro has been added that may be used to apply
1392 tmpfiles configuration during package installation.
1394 * systemd-detect-virt and ConditionVirtualization= now can
1395 detect User-Mode-Linux machines (UML).
1397 * journald will now implicitly log the effective capabilities
1398 set of processes in the message metadata.
1400 * systemd-cryptsetup has gained support for TrueCrypt volumes.
1402 * The initrd interface has been simplified (more specifically,
1403 support for passing performance data via environment
1404 variables and fsck results via files in /run has been
1405 removed). These features were non-essential, and are
1406 nowadays available in a much nicer way by having systemd in
1407 the initrd serialize its state and have the hosts systemd
1408 deserialize it again.
1410 * The udev "keymap" data files and tools to apply keyboard
1411 specific mappings of scan to key codes, and force-release
1412 scan code lists have been entirely replaced by a udev
1413 "keyboard" builtin and a hwdb data file.
1415 * systemd will now honour the kernel's "quiet" command line
1416 argument also during late shutdown, resulting in a
1417 completely silent shutdown when used.
1419 * There's now an option to control the SO_REUSEPORT socket
1420 option in .socket units.
1422 * Instance units will now automatically get a per-template
1423 subslice of system.slice unless something else is explicitly
1424 configured. For example, instances of sshd@.service will now
1425 implicitly be placed in system-sshd.slice rather than
1426 system.slice as before.
1428 * Test coverage support may now be enabled at build time.
1430 Contributions from: Dave Reisner, Frederic Crozat, Harald
1431 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
1432 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
1433 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
1434 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
1435 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
1436 Giokas, Zbigniew Jędrzejewski-Szmek
1438 -- Berlin, 2013-07-23
1442 * Two new unit types have been introduced:
1444 Scope units are very similar to service units, however, are
1445 created out of pre-existing processes -- instead of PID 1
1446 forking off the processes. By using scope units it is
1447 possible for system services and applications to group their
1448 own child processes (worker processes) in a powerful way
1449 which then maybe used to organize them, or kill them
1450 together, or apply resource limits on them.
1452 Slice units may be used to partition system resources in an
1453 hierarchial fashion and then assign other units to them. By
1454 default there are now three slices: system.slice (for all
1455 system services), user.slice (for all user sessions),
1456 machine.slice (for VMs and containers).
1458 Slices and scopes have been introduced primarily in
1459 context of the work to move cgroup handling to a
1460 single-writer scheme, where only PID 1
1461 creates/removes/manages cgroups.
1463 * There's a new concept of "transient" units. In contrast to
1464 normal units these units are created via an API at runtime,
1465 not from configuration from disk. More specifically this
1466 means it is now possible to run arbitrary programs as
1467 independent services, with all execution parameters passed
1468 in via bus APIs rather than read from disk. Transient units
1469 make systemd substantially more dynamic then it ever was,
1470 and useful as a general batch manager.
1472 * logind has been updated to make use of scope and slice units
1473 for managing user sessions. As a user logs in he will get
1474 his own private slice unit, to which all sessions are added
1475 as scope units. We also added support for automatically
1476 adding an instance of user@.service for the user into the
1477 slice. Effectively logind will no longer create cgroup
1478 hierarchies on its own now, it will defer entirely to PID 1
1479 for this by means of scope, service and slice units. Since
1480 user sessions this way become entities managed by PID 1
1481 the output of "systemctl" is now a lot more comprehensive.
1483 * A new mini-daemon "systemd-machined" has been added which
1484 may be used by virtualization managers to register local
1485 VMs/containers. nspawn has been updated accordingly, and
1486 libvirt will be updated shortly. machined will collect a bit
1487 of meta information about the VMs/containers, and assign
1488 them their own scope unit (see above). The collected
1489 meta-data is then made available via the "machinectl" tool,
1490 and exposed in "ps" and similar tools. machined/machinectl
1491 is compile-time optional.
1493 * As discussed earlier, the low-level cgroup configuration
1494 options ControlGroup=, ControlGroupModify=,
1495 ControlGroupPersistent=, ControlGroupAttribute= have been
1496 removed. Please use high-level attribute settings instead as
1497 well as slice units.
1499 * A new bus call SetUnitProperties() has been added to alter
1500 various runtime parameters of a unit. This is primarily
1501 useful to alter cgroup parameters dynamically in a nice way,
1502 but will be extended later on to make more properties
1503 modifiable at runtime. systemctl gained a new set-properties
1504 command that wraps this call.
1506 * A new tool "systemd-run" has been added which can be used to
1507 run arbitrary command lines as transient services or scopes,
1508 while configuring a number of settings via the command
1509 line. This tool is currently very basic, however already
1510 very useful. We plan to extend this tool to even allow
1511 queuing of execution jobs with time triggers from the
1512 command line, similar in fashion to "at".
1514 * nspawn will now inform the user explicitly that kernels with
1515 audit enabled break containers, and suggest the user to turn
1518 * Support for detecting the IMA and AppArmor security
1519 frameworks with ConditionSecurity= has been added.
1521 * journalctl gained a new "-k" switch for showing only kernel
1522 messages, mimicking dmesg output; in addition to "--user"
1523 and "--system" switches for showing only user's own logs
1526 * systemd-delta can now show information about drop-in
1527 snippets extending unit files.
1529 * libsystemd-bus has been substantially updated but is still
1530 not available as public API.
1532 * systemd will now look for the "debug" argument on the kernel
1533 command line and enable debug logging, similar to what
1534 "systemd.log_level=debug" already did before.
1536 * "systemctl set-default", "systemctl get-default" has been
1537 added to configure the default.target symlink, which
1538 controls what to boot into by default.
1540 * "systemctl set-log-level" has been added as a convenient
1541 way to raise and lower systemd logging threshold.
1543 * "systemd-analyze plot" will now show the time the various
1544 generators needed for execution, as well as information
1545 about the unit file loading.
1547 * libsystemd-journal gained a new sd_journal_open_files() call
1548 for opening specific journal files. journactl also gained a
1549 new switch to expose this new functionality. Previously we
1550 only supported opening all files from a directory, or all
1551 files from the system, as opening individual files only is
1552 racy due to journal file rotation.
1554 * systemd gained the new DefaultEnvironment= setting in
1555 /etc/systemd/system.conf to set environment variables for
1558 * If a privileged process logs a journal message with the
1559 OBJECT_PID= field set, then journald will automatically
1560 augment this with additional OBJECT_UID=, OBJECT_GID=,
1561 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
1562 system services want to log events about specific client
1563 processes. journactl/systemctl has been updated to make use
1564 of this information if all log messages regarding a specific
1567 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
1568 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
1569 Reisner, David Coppa, David King, David Strauss, Eelco
1570 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
1571 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
1572 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
1573 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
1574 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
1575 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
1576 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
1577 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
1578 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
1579 Łukasz Stelmach, 장동준
1583 * The Python bindings gained some minimal support for the APIs
1584 exposed by libsystemd-logind.
1586 * ConditionSecurity= gained support for detecting SMACK. Since
1587 this condition already supports SELinux and AppArmor we only
1588 miss IMA for this. Patches welcome!
1590 Contributions from: Karol Lewandowski, Lennart Poettering,
1591 Zbigniew Jędrzejewski-Szmek
1595 * systemd-nspawn will now create /etc/resolv.conf if
1596 necessary, before bind-mounting the host's file onto it.
1598 * systemd-nspawn will now store meta information about a
1599 container on the container's cgroup as extended attribute
1600 fields, including the root directory.
1602 * The cgroup hierarchy has been reworked in many ways. All
1603 objects any of the components systemd creates in the cgroup
1604 tree are now suffixed. More specifically, user sessions are
1605 now placed in cgroups suffixed with ".session", users in
1606 cgroups suffixed with ".user", and nspawn containers in
1607 cgroups suffixed with ".nspawn". Furthermore, all cgroup
1608 names are now escaped in a simple scheme to avoid collision
1609 of userspace object names with kernel filenames. This work
1610 is preparation for making these objects relocatable in the
1611 cgroup tree, in order to allow easy resource partitioning of
1612 these objects without causing naming conflicts.
1614 * systemctl list-dependencies gained the new switches
1615 --plain, --reverse, --after and --before.
1617 * systemd-inhibit now shows the process name of processes that
1618 have taken an inhibitor lock.
1620 * nss-myhostname will now also resolve "localhost"
1621 implicitly. This makes /etc/hosts an optional file and
1622 nicely handles that on IPv6 ::1 maps to both "localhost" and
1625 * libsystemd-logind.so gained a new call
1626 sd_get_machine_names() to enumerate running containers and
1627 VMs (currently only supported by very new libvirt and
1628 nspawn). sd_login_monitor can now be used to watch
1629 VMs/containers coming and going.
1631 * .include is not allowed recursively anymore, and only in
1632 unit files. Usually it is better to use drop-in snippets in
1633 .d/*.conf anyway, as introduced with systemd 198.
1635 * systemd-analyze gained a new "critical-chain" command that
1636 determines the slowest chain of units run during system
1637 boot-up. It is very useful for tracking down where
1638 optimizing boot time is the most beneficial.
1640 * systemd will no longer allow manipulating service paths in
1641 the name=systemd:/system cgroup tree using ControlGroup= in
1642 units. (But is still fine with it in all other dirs.)
1644 * There's a new systemd-nspawn@.service service file that may
1645 be used to easily run nspawn containers as system
1646 services. With the container's root directory in
1647 /var/lib/container/foobar it is now sufficient to run
1648 "systemctl start systemd-nspawn@foobar.service" to boot it.
1650 * systemd-cgls gained a new parameter "--machine" to list only
1651 the processes within a certain container.
1653 * ConditionSecurity= now can check for "apparmor". We still
1654 are lacking checks for SMACK and IMA for this condition
1655 check though. Patches welcome!
1657 * A new configuration file /etc/systemd/sleep.conf has been
1658 added that may be used to configure which kernel operation
1659 systemd is supposed to execute when "suspend", "hibernate"
1660 or "hybrid-sleep" is requested. This makes the new kernel
1661 "freeze" state accessible to the user.
1663 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
1664 the passed argument if applicable.
1666 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
1667 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
1668 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
1669 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
1670 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
1671 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
1672 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
1677 * The output of 'systemctl list-jobs' got some polishing. The
1678 '--type=' argument may now be passed more than once. A new
1679 command 'systemctl list-sockets' has been added which shows
1680 a list of kernel sockets systemd is listening on with the
1681 socket units they belong to, plus the units these socket
1684 * The experimental libsystemd-bus library got substantial
1685 updates to work in conjunction with the (also experimental)
1686 kdbus kernel project. It works well enough to exchange
1687 messages with some sophistication. Note that kdbus is not
1688 ready yet, and the library is mostly an elaborate test case
1689 for now, and not installable.
1691 * systemd gained a new unit 'systemd-static-nodes.service'
1692 that generates static device nodes earlier during boot, and
1693 can run in conjunction with udev.
1695 * libsystemd-login gained a new call sd_pid_get_user_unit()
1696 to retrieve the user systemd unit a process is running
1697 in. This is useful for systems where systemd is used as
1700 * systemd-nspawn now places all containers in the new /machine
1701 top-level cgroup directory in the name=systemd
1702 hierarchy. libvirt will soon do the same, so that we get a
1703 uniform separation of /system, /user and /machine for system
1704 services, user processes and containers/virtual
1705 machines. This new cgroup hierarchy is also useful to stick
1706 stable names to specific container instances, which can be
1707 recognized later this way (this name may be controlled
1708 via systemd-nspawn's new -M switch). libsystemd-login also
1709 gained a new call sd_pid_get_machine_name() to retrieve the
1710 name of the container/VM a specific process belongs to.
1712 * bootchart can now store its data in the journal.
1714 * libsystemd-journal gained a new call
1715 sd_journal_add_conjunction() for AND expressions to the
1716 matching logic. This can be used to express more complex
1717 logical expressions.
1719 * journactl can now take multiple --unit= and --user-unit=
1722 * The cryptsetup logic now understands the "luks.key=" kernel
1723 command line switch for specifying a file to read the
1724 decryption key from. Also, if a configured key file is not
1725 found the tool will now automatically fall back to prompting
1728 * Python systemd.journal module was updated to wrap recently
1729 added functions from libsystemd-journal. The interface was
1730 changed to bring the low level interface in s.j._Reader
1731 closer to the C API, and the high level interface in
1732 s.j.Reader was updated to wrap and convert all data about
1735 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
1736 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
1737 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
1738 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
1739 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
1740 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1744 * journalctl --update-catalog now understands a new --root=
1745 option to operate on catalogs found in a different root
1748 * During shutdown after systemd has terminated all running
1749 services a final killing loop kills all remaining left-over
1750 processes. We will now print the name of these processes
1751 when we send SIGKILL to them, since this usually indicates a
1754 * If /etc/crypttab refers to password files stored on
1755 configured mount points automatic dependencies will now be
1756 generated to ensure the specific mount is established first
1757 before the key file is attempted to be read.
1759 * 'systemctl status' will now show information about the
1760 network sockets a socket unit is listening on.
1762 * 'systemctl status' will also shown information about any
1763 drop-in configuration file for units. (Drop-In configuration
1764 files in this context are files such as
1765 /etc/systemd/systemd/foobar.service.d/*.conf)
1767 * systemd-cgtop now optionally shows summed up CPU times of
1768 cgroups. Press '%' while running cgtop to switch between
1769 percentage and absolute mode. This is useful to determine
1770 which cgroups use up the most CPU time over the entire
1771 runtime of the system. systemd-cgtop has also been updated
1772 to be 'pipeable' for processing with further shell tools.
1774 * 'hostnamectl set-hostname' will now allow setting of FQDN
1777 * The formatting and parsing of time span values has been
1778 changed. The parser now understands fractional expressions
1779 such as "5.5h". The formatter will now output fractional
1780 expressions for all time spans under 1min, i.e. "5.123456s"
1781 rather than "5s 123ms 456us". For time spans under 1s
1782 millisecond values are shown, for those under 1ms
1783 microsecond values are shown. This should greatly improve
1784 all time-related output of systemd.
1786 * libsystemd-login and libsystemd-journal gained new
1787 functions for querying the poll() events mask and poll()
1788 timeout value for integration into arbitrary event
1791 * localectl gained the ability to list available X11 keymaps
1792 (models, layouts, variants, options).
1794 * 'systemd-analyze dot' gained the ability to filter for
1795 specific units via shell-style globs, to create smaller,
1796 more useful graphs. I.e. it is now possible to create simple
1797 graphs of all the dependencies between only target units, or
1798 of all units that Avahi has dependencies with.
1800 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
1801 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
1802 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
1803 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
1804 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
1805 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
1806 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
1810 * The boot-time readahead implementation for rotating media
1811 will now read the read-ahead data in multiple passes which
1812 consist of all read requests made in equidistant time
1813 intervals. This means instead of strictly reading read-ahead
1814 data in its physical order on disk we now try to find a
1815 middle ground between physical and access time order.
1817 * /etc/os-release files gained a new BUILD_ID= field for usage
1818 on operating systems that provide continuous builds of OS
1821 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
1822 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
1823 William Douglas, Zbigniew Jędrzejewski-Szmek
1827 * systemd-python gained an API exposing libsystemd-daemon.
1829 * The SMACK setup logic gained support for uploading CIPSO
1832 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
1833 ReadOnlyDirectories= and InaccessibleDirectories= has
1834 changed. The private /tmp and /var/tmp directories are now
1835 shared by all processes of a service (which means
1836 ExecStartPre= may now leave data in /tmp that ExecStart= of
1837 the same service can still access). When a service is
1838 stopped its temporary directories are immediately deleted
1839 (normal clean-up with tmpfiles is still done in addition to
1842 * By default, systemd will now set a couple of sysctl
1843 variables in the kernel: the safe sysrq options are turned
1844 on, IP route verification is turned on, and source routing
1845 disabled. The recently added hardlink and softlink
1846 protection of the kernel is turned on. These settings should
1847 be reasonably safe, and good defaults for all new systems.
1849 * The predictable network naming logic may now be turned off
1850 with a new kernel command line switch: net.ifnames=0.
1852 * A new libsystemd-bus module has been added that implements a
1853 pretty complete D-Bus client library. For details see:
1855 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
1857 * journald will now explicitly flush the journal files to disk
1858 at the latest 5min after each write. The file will then also
1859 be marked offline until the next write. This should increase
1860 reliability in case of a crash. The synchronization delay
1861 can be configured via SyncIntervalSec= in journald.conf.
1863 * There's a new remote-fs-setup.target unit that can be used
1864 to pull in specific services when at least one remote file
1865 system is to be mounted.
1867 * There are new targets timers.target and paths.target as
1868 canonical targets to pull user timer and path units in
1869 from. This complements sockets.target with a similar
1870 purpose for socket units.
1872 * libudev gained a new call udev_device_set_attribute_value()
1873 to set sysfs attributes of a device.
1875 * The udev daemon now sets the default number of worker
1876 processes executed in parallel based on the number of available
1877 CPUs instead of the amount of available RAM. This is supposed
1878 to provide a more reliable default and limit a too aggressive
1879 paralellism for setups with 1000s of devices connected.
1881 Contributions from: Auke Kok, Colin Walters, Cristian
1882 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
1883 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
1884 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
1885 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
1886 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
1887 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
1888 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
1889 Zbigniew Jędrzejewski-Szmek
1893 * Configuration of unit files may now be extended via drop-in
1894 files without having to edit/override the unit files
1895 themselves. More specifically, if the administrator wants to
1896 change one value for a service file foobar.service he can
1897 now do so by dropping in a configuration snippet into
1898 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
1899 will load all these snippets and apply them on top of the
1900 main unit configuration file, possibly extending or
1901 overriding its settings. Using these drop-in snippets is
1902 generally nicer than the two earlier options for changing
1903 unit files locally: copying the files from
1904 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
1905 them there; or creating a new file in /etc/systemd/system/
1906 that incorporates the original one via ".include". Drop-in
1907 snippets into these .d/ directories can be placed in any
1908 directory systemd looks for units in, and the usual
1909 overriding semantics between /usr/lib, /etc and /run apply
1912 * Most unit file settings which take lists of items can now be
1913 reset by assigning the empty string to them. For example,
1914 normally, settings such as Environment=FOO=BAR append a new
1915 environment variable assignment to the environment block,
1916 each time they are used. By assigning Environment= the empty
1917 string the environment block can be reset to empty. This is
1918 particularly useful with the .d/*.conf drop-in snippets
1919 mentioned above, since this adds the ability to reset list
1920 settings from vendor unit files via these drop-ins.
1922 * systemctl gained a new "list-dependencies" command for
1923 listing the dependencies of a unit recursively.
1925 * Inhibitors are now honored and listed by "systemctl
1926 suspend", "systemctl poweroff" (and similar) too, not only
1927 GNOME. These commands will also list active sessions by
1930 * Resource limits (as exposed by the various control group
1931 controllers) can now be controlled dynamically at runtime
1932 for all units. More specifically, you can now use a command
1933 like "systemctl set-cgroup-attr foobar.service cpu.shares
1934 2000" to alter the CPU shares a specific service gets. These
1935 settings are stored persistently on disk, and thus allow the
1936 administrator to easily adjust the resource usage of
1937 services with a few simple commands. This dynamic resource
1938 management logic is also available to other programs via the
1939 bus. Almost any kernel cgroup attribute and controller is
1942 * systemd-vconsole-setup will now copy all font settings to
1943 all allocated VTs, where it previously applied them only to
1946 * libsystemd-login gained the new sd_session_get_tty() API
1949 * This release drops support for a few legacy or
1950 distribution-specific LSB facility names when parsing init
1951 scripts: $x-display-manager, $mail-transfer-agent,
1952 $mail-transport-agent, $mail-transfer-agent, $smtp,
1953 $null. Also, the mail-transfer-agent.target unit backing
1954 this has been removed. Distributions which want to retain
1955 compatibility with this should carry the burden for
1956 supporting this themselves and patch support for these back
1957 in, if they really need to. Also, the facilities $syslog and
1958 $local_fs are now ignored, since systemd does not support
1959 early-boot LSB init scripts anymore, and these facilities
1960 are implied anyway for normal services. syslog.target has
1963 * There are new bus calls on PID1's Manager object for
1964 cancelling jobs, and removing snapshot units. Previously,
1965 both calls were only available on the Job and Snapshot
1968 * systemd-journal-gatewayd gained SSL support.
1970 * The various "environment" files, such as /etc/locale.conf
1971 now support continuation lines with a backslash ("\") as
1972 last character in the line, similarly in style (but different)
1973 to how this is supported in shells.
1975 * For normal user processes the _SYSTEMD_USER_UNIT= field is
1976 now implicitly appended to every log entry logged. systemctl
1977 has been updated to filter by this field when operating on a
1978 user systemd instance.
1980 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
1981 CAP_AUDIT_CONTROL capabilities to the capabilities set for
1982 the container. This makes it easier to boot unmodified
1983 Fedora systems in a container, which however still requires
1984 audit=0 to be passed on the kernel command line. Auditing in
1985 kernel and userspace is unfortunately still too broken in
1986 context of containers, hence we recommend compiling it out
1987 of the kernel or using audit=0. Hopefully this will be fixed
1988 one day for good in the kernel.
1990 * nspawn gained the new --bind= and --bind-ro= parameters to
1991 bind mount specific directories from the host into the
1994 * nspawn will now mount its own devpts file system instance
1995 into the container, in order not to leak pty devices from
1996 the host into the container.
1998 * systemd will now read the firmware boot time performance
1999 information from the EFI variables, if the used boot loader
2000 supports this, and takes it into account for boot performance
2001 analysis via "systemd-analyze". This is currently supported
2002 only in conjunction with Gummiboot, but could be supported
2003 by other boot loaders too. For details see:
2005 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2007 * A new generator has been added that automatically mounts the
2008 EFI System Partition (ESP) to /boot, if that directory
2009 exists, is empty, and no other file system has been
2010 configured to be mounted there.
2012 * logind will now send out PrepareForSleep(false) out
2013 unconditionally, after coming back from suspend. This may be
2014 used by applications as asynchronous notification for
2015 system resume events.
2017 * "systemctl unlock-sessions" has been added, that allows
2018 unlocking the screens of all user sessions at once, similar
2019 to how "systemctl lock-sessions" already locked all users
2020 sessions. This is backed by a new D-Bus call UnlockSessions().
2022 * "loginctl seat-status" will now show the master device of a
2023 seat. (i.e. the device of a seat that needs to be around for
2024 the seat to be considered available, usually the graphics
2027 * tmpfiles gained a new "X" line type, that allows
2028 configuration of files and directories (with wildcards) that
2029 shall be excluded from automatic cleanup ("aging").
2031 * udev default rules set the device node permissions now only
2032 at "add" events, and do not change them any longer with a
2033 later "change" event.
2035 * The log messages for lid events and power/sleep keypresses
2036 now carry a message ID.
2038 * We now have a substantially larger unit test suite, but this
2039 continues to be work in progress.
2041 * udevadm hwdb gained a new --root= parameter to change the
2042 root directory to operate relative to.
2044 * logind will now issue a background sync() request to the kernel
2045 early at shutdown, so that dirty buffers are flushed to disk early
2046 instead of at the last moment, in order to optimize shutdown
2049 * A new bootctl tool has been added that is an interface for
2050 certain boot loader operations. This is currently a preview
2051 and is likely to be extended into a small mechanism daemon
2052 like timedated, localed, hostnamed, and can be used by
2053 graphical UIs to enumerate available boot options, and
2054 request boot into firmware operations.
2056 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2057 the rest of the package. It also has been updated to work
2058 correctly in initrds.
2060 * Policykit previously has been runtime optional, and is now
2061 also compile time optional via a configure switch.
2063 * systemd-analyze has been reimplemented in C. Also "systemctl
2064 dot" has moved into systemd-analyze.
2066 * "systemctl status" with no further parameters will now print
2067 the status of all active or failed units.
2069 * Operations such as "systemctl start" can now be executed
2070 with a new mode "--irreversible" which may be used to queue
2071 operations that cannot accidentally be reversed by a later
2072 job queuing. This is by default used to make shutdown
2073 requests more robust.
2075 * The Python API of systemd now gained a new module for
2076 reading journal files.
2078 * A new tool kernel-install has been added that can install
2079 kernel images according to the Boot Loader Specification:
2081 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2083 * Boot time console output has been improved to provide
2084 animated boot time output for hanging jobs.
2086 * A new tool systemd-activate has been added which can be used
2087 to test socket activation with, directly from the command
2088 line. This should make it much easier to test and debug
2089 socket activation in daemons.
2091 * journalctl gained a new "--reverse" (or -r) option to show
2092 journal output in reverse order (i.e. newest line first).
2094 * journalctl gained a new "--pager-end" (or -e) option to jump
2095 to immediately jump to the end of the journal in the
2096 pager. This is only supported in conjunction with "less".
2098 * journalctl gained a new "--user-unit=" option, that works
2099 similarly to "--unit=" but filters for user units rather than
2102 * A number of unit files to ease adoption of systemd in
2103 initrds has been added. This moves some minimal logic from
2104 the various initrd implementations into systemd proper.
2106 * The journal files are now owned by a new group
2107 "systemd-journal", which exists specifically to allow access
2108 to the journal, and nothing else. Previously, we used the
2109 "adm" group for that, which however possibly covers more
2110 than just journal/log file access. This new group is now
2111 already used by systemd-journal-gatewayd to ensure this
2112 daemon gets access to the journal files and as little else
2113 as possible. Note that "make install" will also set FS ACLs
2114 up for /var/log/journal to give "adm" and "wheel" read
2115 access to it, in addition to "systemd-journal" which owns
2116 the journal files. We recommend that packaging scripts also
2117 add read access to "adm" + "wheel" to /var/log/journal, and
2118 all existing/future journal files. To normal users and
2119 administrators little changes, however packagers need to
2120 ensure to create the "systemd-journal" system group at
2121 package installation time.
2123 * The systemd-journal-gatewayd now runs as unprivileged user
2124 systemd-journal-gateway:systemd-journal-gateway. Packaging
2125 scripts need to create these system user/group at
2128 * timedated now exposes a new boolean property CanNTP that
2129 indicates whether a local NTP service is available or not.
2131 * systemd-detect-virt will now also detect xen PVs
2133 * The pstore file system is now mounted by default, if it is
2136 * In addition to the SELinux and IMA policies we will now also
2137 load SMACK policies at early boot.
2139 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2140 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
2141 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
2142 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
2143 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
2144 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
2145 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
2146 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
2147 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
2148 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
2149 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
2150 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
2151 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
2152 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
2156 * Timer units now support calendar time events in addition to
2157 monotonic time events. That means you can now trigger a unit
2158 based on a calendar time specification such as "Thu,Fri
2159 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
2160 or fifth day of any month of the year 2013, given that it is
2161 a thursday or friday. This brings timer event support
2162 considerably closer to cron's capabilities. For details on
2163 the supported calendar time specification language see
2166 * udev now supports a number of different naming policies for
2167 network interfaces for predictable names, and a combination
2168 of these policies is now the default. Please see this wiki
2169 document for details:
2171 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
2173 * Auke Kok's bootchart implementation has been added to the
2174 systemd tree. It is an optional component that can graph the
2175 boot in quite some detail. It is one of the best bootchart
2176 implementations around and minimal in its code and
2179 * nss-myhostname has been integrated into the systemd source
2180 tree. nss-myhostname guarantees that the local hostname
2181 always stays resolvable via NSS. It has been a weak
2182 requirement of systemd-hostnamed since a long time, and
2183 since its code is actually trivial we decided to just
2184 include it in systemd's source tree. It can be turned off
2185 with a configure switch.
2187 * The read-ahead logic is now capable of properly detecting
2188 whether a btrfs file system is on SSD or rotating media, in
2189 order to optimize the read-ahead scheme. Previously, it was
2190 only capable of detecting this on traditional file systems
2193 * In udev, additional device properties are now read from the
2194 IAB in addition to the OUI database. Also, Bluetooth company
2195 identities are attached to the devices as well.
2197 * In service files %U may be used as specifier that is
2198 replaced by the configured user name of the service.
2200 * nspawn may now be invoked without a controlling TTY. This
2201 makes it suitable for invocation as its own service. This
2202 may be used to set up a simple containerized server system
2203 using only core OS tools.
2205 * systemd and nspawn can now accept socket file descriptors
2206 when they are started for socket activation. This enables
2207 implementation of socket activated nspawn
2208 containers. i.e. think about autospawning an entire OS image
2209 when the first SSH or HTTP connection is received. We expect
2210 that similar functionality will also be added to libvirt-lxc
2213 * journalctl will now suppress ANSI color codes when
2214 presenting log data.
2216 * systemctl will no longer show control group information for
2217 a unit if a the control group is empty anyway.
2219 * logind can now automatically suspend/hibernate/shutdown the
2222 * /etc/machine-info and hostnamed now also expose the chassis
2223 type of the system. This can be used to determine whether
2224 the local system is a laptop, desktop, handset or
2225 tablet. This information may either be configured by the
2226 user/vendor or is automatically determined from ACPI and DMI
2227 information if possible.
2229 * A number of PolicyKit actions are now bound together with
2230 "imply" rules. This should simplify creating UIs because
2231 many actions will now authenticate similar ones as well.
2233 * Unit files learnt a new condition ConditionACPower= which
2234 may be used to conditionalize a unit depending on whether an
2235 AC power source is connected or not, of whether the system
2236 is running on battery power.
2238 * systemctl gained a new "is-failed" verb that may be used in
2239 shell scripts and suchlike to check whether a specific unit
2240 is in the "failed" state.
2242 * The EnvironmentFile= setting in unit files now supports file
2243 globbing, and can hence be used to easily read a number of
2244 environment files at once.
2246 * systemd will no longer detect and recognize specific
2247 distributions. All distribution-specific #ifdeffery has been
2248 removed, systemd is now fully generic and
2249 distribution-agnostic. Effectively, not too much is lost as
2250 a lot of the code is still accessible via explicit configure
2251 switches. However, support for some distribution specific
2252 legacy configuration file formats has been dropped. We
2253 recommend distributions to simply adopt the configuration
2254 files everybody else uses now and convert the old
2255 configuration from packaging scripts. Most distributions
2256 already did that. If that's not possible or desirable,
2257 distributions are welcome to forward port the specific
2258 pieces of code locally from the git history.
2260 * When logging a message about a unit systemd will now always
2261 log the unit name in the message meta data.
2263 * localectl will now also discover system locale data that is
2264 not stored in locale archives, but directly unpacked.
2266 * logind will no longer unconditionally use framebuffer
2267 devices as seat masters, i.e. as devices that are required
2268 to be existing before a seat is considered preset. Instead,
2269 it will now look for all devices that are tagged as
2270 "seat-master" in udev. By default framebuffer devices will
2271 be marked as such, but depending on local systems other
2272 devices might be marked as well. This may be used to
2273 integrate graphics cards using closed source drivers (such
2274 as NVidia ones) more nicely into logind. Note however, that
2275 we recommend using the open source NVidia drivers instead,
2276 and no udev rules for the closed-source drivers will be
2277 shipped from us upstream.
2279 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2280 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2281 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2282 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2283 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2284 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2285 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2286 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2287 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2288 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2289 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2290 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2295 * udev gained support for loading additional device properties
2296 from an indexed database that is keyed by vendor/product IDs
2297 and similar device identifiers. For the beginning this
2298 "hwdb" is populated with data from the well-known PCI and
2299 USB database, but also includes PNP, ACPI and OID data. In
2300 the longer run this indexed database shall grow into
2301 becoming the one central database for non-essential
2302 userspace device metadata. Previously, data from the PCI/USB
2303 database was only attached to select devices, since the
2304 lookup was a relatively expensive operation due to O(n) time
2305 complexity (with n being the number of entries in the
2306 database). Since this is now O(1), we decided to add in this
2307 data for all devices where this is available, by
2308 default. Note that the indexed database needs to be rebuilt
2309 when new data files are installed. To achieve this you need
2310 to update your packaging scripts to invoke "udevadm hwdb
2311 --update" after installation of hwdb data files. For
2312 RPM-based distributions we introduced the new
2313 %udev_hwdb_update macro for this purpose.
2315 * The Journal gained support for the "Message Catalog", an
2316 indexed database to link up additional information with
2317 journal entries. For further details please check:
2319 http://www.freedesktop.org/wiki/Software/systemd/catalog
2321 The indexed message catalog database also needs to be
2322 rebuilt after installation of message catalog files. Use
2323 "journalctl --update-catalog" for this. For RPM-based
2324 distributions we introduced the %journal_catalog_update
2325 macro for this purpose.
2327 * The Python Journal bindings gained support for the standard
2328 Python logging framework.
2330 * The Journal API gained new functions for checking whether
2331 the underlying file system of a journal file is capable of
2332 properly reporting file change notifications, or whether
2333 applications that want to reflect journal changes "live"
2334 need to recheck journal files continuously in appropriate
2337 * It is now possible to set the "age" field for tmpfiles
2338 entries to 0, indicating that files matching this entry
2339 shall always be removed when the directories are cleaned up.
2341 * coredumpctl gained a new "gdb" verb which invokes gdb
2342 right-away on the selected coredump.
2344 * There's now support for "hybrid sleep" on kernels that
2345 support this, in addition to "suspend" and "hibernate". Use
2346 "systemctl hybrid-sleep" to make use of this.
2348 * logind's HandleSuspendKey= setting (and related settings)
2349 now gained support for a new "lock" setting to simply
2350 request the screen lock on all local sessions, instead of
2351 actually executing a suspend or hibernation.
2353 * systemd will now mount the EFI variables file system by
2356 * Socket units now gained support for configuration of the
2357 SMACK security label.
2359 * timedatectl will now output the time of the last and next
2360 daylight saving change.
2362 * We dropped support for various legacy and distro-specific
2363 concepts, such as insserv, early-boot SysV services
2364 (i.e. those for non-standard runlevels such as 'b' or 'S')
2365 or ArchLinux /etc/rc.conf support. We recommend the
2366 distributions who still need support this to either continue
2367 to maintain the necessary patches downstream, or find a
2368 different solution. (Talk to us if you have questions!)
2370 * Various systemd components will now bypass PolicyKit checks
2371 for root and otherwise handle properly if PolicyKit is not
2372 found to be around. This should fix most issues for
2373 PolicyKit-less systems. Quite frankly this should have been
2374 this way since day one. It is absolutely our intention to
2375 make systemd work fine on PolicyKit-less systems, and we
2376 consider it a bug if something does not work as it should if
2377 PolicyKit is not around.
2379 * For embedded systems it is now possible to build udev and
2380 systemd without blkid and/or kmod support.
2382 * "systemctl switch-root" is now capable of switching root
2383 more than once. I.e. in addition to transitions from the
2384 initrd to the host OS it is now possible to transition to
2385 further OS images from the host. This is useful to implement
2386 offline updating tools.
2388 * Various other additions have been made to the RPM macros
2389 shipped with systemd. Use %udev_rules_update() after
2390 installing new udev rules files. %_udevhwdbdir,
2391 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
2392 %_sysctldir are now available which resolve to the right
2393 directories for packages to place various data files in.
2395 * journalctl gained the new --full switch (in addition to
2396 --all, to disable ellipsation for long messages.
2398 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
2399 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
2400 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
2401 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
2402 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
2403 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
2404 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
2405 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
2406 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
2410 * journalctl gained new --since= and --until= switches to
2411 filter by time. It also now supports nice filtering for
2412 units via --unit=/-u.
2414 * Type=oneshot services may use ExecReload= and do the
2417 * The journal daemon now supports time-based rotation and
2418 vacuuming, in addition to the usual disk-space based
2421 * The journal will now index the available field values for
2422 each field name. This enables clients to show pretty drop
2423 downs of available match values when filtering. The bash
2424 completion of journalctl has been updated
2425 accordingly. journalctl gained a new switch -F to list all
2426 values a certain field takes in the journal database.
2428 * More service events are now written as structured messages
2429 to the journal, and made recognizable via message IDs.
2431 * The timedated, localed and hostnamed mini-services which
2432 previously only provided support for changing time, locale
2433 and hostname settings from graphical DEs such as GNOME now
2434 also have a minimal (but very useful) text-based client
2435 utility each. This is probably the nicest way to changing
2436 these settings from the command line now, especially since
2437 it lists available options and is fully integrated with bash
2440 * There's now a new tool "systemd-coredumpctl" to list and
2441 extract coredumps from the journal.
2443 * We now install a README each in /var/log/ and
2444 /etc/rc.d/init.d explaining where the system logs and init
2445 scripts went. This hopefully should help folks who go to
2446 that dirs and look into the otherwise now empty void and
2447 scratch their heads.
2449 * When user-services are invoked (by systemd --user) the
2450 $MANAGERPID env var is set to the PID of systemd.
2452 * SIGRTMIN+24 when sent to a --user instance will now result
2453 in immediate termination of systemd.
2455 * gatewayd received numerous feature additions such as a
2456 "follow" mode, for live syncing and filtering.
2458 * browse.html now allows filtering and showing detailed
2459 information on specific entries. Keyboard navigation and
2460 mouse screen support has been added.
2462 * gatewayd/journalctl now supports HTML5/JSON
2463 Server-Sent-Events as output.
2465 * The SysV init script compatibility logic will now
2466 heuristically determine whether a script supports the
2467 "reload" verb, and only then make this available as
2470 * "systemctl status --follow" has been removed, use "journalctl
2473 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
2474 have been removed since they are hardly useful to be
2477 * And I'd like to take the opportunity to specifically mention
2478 Zbigniew for his great contributions. Zbigniew, you rock!
2480 Contributions from: Andrew Eikum, Christian Hesse, Colin
2481 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
2482 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
2483 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
2484 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
2485 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
2486 Jędrzejewski-Szmek, Сковорода Никита Андреевич
2490 * If /etc/vconsole.conf is non-existent or empty we will no
2491 longer load any console font or key map at boot by
2492 default. Instead the kernel defaults will be left
2493 intact. This is definitely the right thing to do, as no
2494 configuration should mean no configuration, and hard-coding
2495 font names that are different on all archs is probably a bad
2496 idea. Also, the kernel default key map and font should be
2497 good enough for most cases anyway, and mostly identical to
2498 the userspace fonts/key maps we previously overloaded them
2499 with. If distributions want to continue to default to a
2500 non-kernel font or key map they should ship a default
2501 /etc/vconsole.conf with the appropriate contents.
2503 Contributions from: Colin Walters, Daniel J Walsh, Dave
2504 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
2505 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2509 * journalctl gained a new --cursor= switch to show entries
2510 starting from the specified location in the journal.
2512 * We now enforce a size limit on journal entry fields exported
2513 with "-o json" in journalctl. Fields larger than 4K will be
2514 assigned null. This can be turned off with --all.
2516 * An (optional) journal gateway daemon is now available as
2517 "systemd-journal-gatewayd.service". This service provides
2518 access to the journal via HTTP and JSON. This functionality
2519 will be used to implement live log synchronization in both
2520 pull and push modes, but has various other users too, such
2521 as easy log access for debugging of embedded devices. Right
2522 now it is already useful to retrieve the journal via HTTP:
2524 # systemctl start systemd-journal-gatewayd.service
2525 # wget http://localhost:19531/entries
2527 This will download the journal contents in a
2528 /var/log/messages compatible format. The same as JSON:
2530 # curl -H"Accept: application/json" http://localhost:19531/entries
2532 This service is also accessible via a web browser where a
2533 single static HTML5 app is served that uses the JSON logic
2534 to enable the user to do some basic browsing of the
2535 journal. This will be extended later on. Here's an example
2536 screenshot of this app in its current state:
2538 http://0pointer.de/public/journal-gatewayd
2540 Contributions from: Kay Sievers, Lennart Poettering, Robert
2541 Milasan, Tom Gundersen
2545 * The bash completion logic is now available for journalctl
2548 * We do not mount the "cpuset" controller anymore together with
2549 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
2550 started if no parameters are assigned to it. "cpuset" hence
2551 broke code that assumed it it could create "cpu" groups and
2554 * journalctl -f will now subscribe to terminal size changes,
2555 and line break accordingly.
2557 Contributions from: Dave Reisner, Kay Sievers, Lennart
2558 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
2562 * nspawn will now create a symlink /etc/localtime in the
2563 container environment, copying the host's timezone
2564 setting. Previously this has been done via a bind mount, but
2565 since symlinks cannot be bind mounted this has now been
2566 changed to create/update the appropriate symlink.
2568 * journalctl -n's line number argument is now optional, and
2569 will default to 10 if omitted.
2571 * journald will now log the maximum size the journal files may
2572 take up on disk. This is particularly useful if the default
2573 built-in logic of determining this parameter from the file
2574 system size is used. Use "systemctl status
2575 systemd-journald.service" to see this information.
2577 * The multi-seat X wrapper tool has been stripped down. As X
2578 is now capable of enumerating graphics devices via udev in a
2579 seat-aware way the wrapper is not strictly necessary
2580 anymore. A stripped down temporary stop-gap is still shipped
2581 until the upstream display managers have been updated to
2582 fully support the new X logic. Expect this wrapper to be
2583 removed entirely in one of the next releases.
2585 * HandleSleepKey= in logind.conf has been split up into
2586 HandleSuspendKey= and HandleHibernateKey=. The old setting
2587 is not available anymore. X11 and the kernel are
2588 distuingishing between these keys and we should too. This
2589 also means the inhibition lock for these keys has been split
2592 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
2593 Poettering, Lukas Nykryn, Václav Pavlín
2597 * Whenever a unit changes state we will now log this to the
2598 journal and show along the unit's own log output in
2601 * ConditionPathIsMountPoint= can now properly detect bind
2602 mount points too. (Previously, a bind mount of one file
2603 system to another place in the same file system could not be
2604 detected as mount, since they shared struct stat's st_dev
2607 * We will now mount the cgroup controllers cpu, cpuacct,
2608 cpuset and the controllers net_cls, net_prio together by
2611 * nspawn containers will now have a virtualized boot
2612 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
2613 over with a randomized ID at container initialization). This
2614 has the effect of making "journalctl -b" do the right thing
2617 * The JSON output journal serialization has been updated not
2618 to generate "endless" list objects anymore, but rather one
2619 JSON object per line. This is more in line how most JSON
2620 parsers expect JSON objects. The new output mode
2621 "json-pretty" has been added to provide similar output, but
2622 neatly aligned for readability by humans.
2624 * We dropped all explicit sync() invocations in the shutdown
2625 code. The kernel does this implicitly anyway in the kernel
2626 reboot() syscall. halt(8)'s -n option is now a compatibility
2629 * We now support virtualized reboot() in containers, as
2630 supported by newer kernels. We will fall back to exit() if
2631 CAP_SYS_REBOOT is not available to the container. Also,
2632 nspawn makes use of this now and will actually reboot the
2633 container if the containerized OS asks for that.
2635 * journalctl will only show local log output by default
2636 now. Use --merge (-m) to show remote log output, too.
2638 * libsystemd-journal gained the new sd_journal_get_usage()
2639 call to determine the current disk usage of all journal
2640 files. This is exposed in the new "journalctl --disk-usage"
2643 * journald gained a new configuration setting SplitMode= in
2644 journald.conf which may be used to control how user journals
2645 are split off. See journald.conf(5) for details.
2647 * A new condition type ConditionFileNotEmpty= has been added.
2649 * tmpfiles' "w" lines now support file globbing, to write
2650 multiple files at once.
2652 * We added Python bindings for the journal submission
2653 APIs. More Python APIs for a number of selected APIs will
2654 likely follow. Note that we intend to add native bindings
2655 only for the Python language, as we consider it common
2656 enough to deserve bindings shipped within systemd. There are
2657 various projects outside of systemd that provide bindings
2658 for languages such as PHP or Lua.
2660 * Many conditions will now resolve specifiers such as %i. In
2661 addition, PathChanged= and related directives of .path units
2662 now support specifiers as well.
2664 * There's now a new RPM macro definition for the system preset
2667 * journald will now warn if it ca not forward a message to the
2668 syslog daemon because its socket is full.
2670 * timedated will no longer write or process /etc/timezone,
2671 except on Debian. As we do not support late mounted /usr
2672 anymore /etc/localtime always being a symlink is now safe,
2673 and hence the information in /etc/timezone is not necessary
2676 * logind will now always reserve one VT for a text getty (VT6
2677 by default). Previously if more than 6 X sessions where
2678 started they took up all the VTs with auto-spawned gettys,
2679 so that no text gettys were available anymore.
2681 * udev will now automatically inform the btrfs kernel logic
2682 about btrfs RAID components showing up. This should make
2683 simple hotplug based btrfs RAID assembly work.
2685 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
2686 (but not for its children which will stay at the kernel
2687 default). This should allow setups with a lot more listening
2690 * systemd will now always pass the configured timezone to the
2691 kernel at boot. timedated will do the same when the timezone
2694 * logind's inhibition logic has been updated. By default,
2695 logind will now handle the lid switch, the power and sleep
2696 keys all the time, even in graphical sessions. If DEs want
2697 to handle these events on their own they should take the new
2698 handle-power-key, handle-sleep-key and handle-lid-switch
2699 inhibitors during their runtime. A simple way to achiveve
2700 that is to invoke the DE wrapped in an invocation of:
2702 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
2704 * Access to unit operations is now checked via SELinux taking
2705 the unit file label and client process label into account.
2707 * systemd will now notify the administrator in the journal
2708 when he over-mounts a non-empty directory.
2710 * There are new specifiers that are resolved in unit files,
2711 for the host name (%H), the machine ID (%m) and the boot ID
2714 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
2715 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
2716 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
2717 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
2718 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
2719 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
2720 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
2724 * Support for reading structured kernel messages from
2725 /dev/kmsg has now been added and is enabled by default.
2727 * Support for reading kernel messages from /proc/kmsg has now
2728 been removed. If you want kernel messages in the journal
2729 make sure to run a recent kernel (>= 3.5) that supports
2730 reading structured messages from /dev/kmsg (see
2731 above). /proc/kmsg is now exclusive property of classic
2732 syslog daemons again.
2734 * The libudev API gained the new
2735 udev_device_new_from_device_id() call.
2737 * The logic for file system namespace (ReadOnlyDirectory=,
2738 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
2739 require pivot_root() anymore. This means fewer temporary
2740 directories are created below /tmp for this feature.
2742 * nspawn containers will now see and receive all submounts
2743 made on the host OS below the root file system of the
2746 * Forward Secure Sealing is now supported for Journal files,
2747 which provide cryptographical sealing of journal files so
2748 that attackers cannot alter log history anymore without this
2749 being detectable. Lennart will soon post a blog story about
2750 this explaining it in more detail.
2752 * There are two new service settings RestartPreventExitStatus=
2753 and SuccessExitStatus= which allow configuration of exit
2754 status (exit code or signal) which will be excepted from the
2755 restart logic, resp. consider successful.
2757 * journalctl gained the new --verify switch that can be used
2758 to check the integrity of the structure of journal files and
2759 (if Forward Secure Sealing is enabled) the contents of
2762 * nspawn containers will now be run with /dev/stdin, /dev/fd/
2763 and similar symlinks pre-created. This makes running shells
2764 as container init process a lot more fun.
2766 * The fstab support can now handle PARTUUID= and PARTLABEL=
2769 * A new ConditionHost= condition has been added to match
2770 against the hostname (with globs) and machine ID. This is
2771 useful for clusters where a single OS image is used to
2772 provision a large number of hosts which shall run slightly
2773 different sets of services.
2775 * Services which hit the restart limit will now be placed in a
2778 Contributions from: Bertram Poettering, Dave Reisner, Huang
2779 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
2780 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
2784 * When running in --user mode systemd will now become a
2785 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
2786 tree a lot more organized.
2788 * A new PartOf= unit dependency type has been introduced that
2789 may be used to group services in a natural way.
2791 * "systemctl enable" may now be used to enable instances of
2794 * journalctl now prints error log levels in red, and
2795 warning/notice log levels in bright white. It also supports
2796 filtering by log level now.
2798 * cgtop gained a new -n switch (similar to top), to configure
2799 the maximum number of iterations to run for. It also gained
2800 -b, to run in batch mode (accepting no input).
2802 * The suffix ".service" may now be omitted on most systemctl
2803 command lines involving service unit names.
2805 * There's a new bus call in logind to lock all sessions, as
2806 well as a loginctl verb for it "lock-sessions".
2808 * libsystemd-logind.so gained a new call sd_journal_perror()
2809 that works similar to libc perror() but logs to the journal
2810 and encodes structured information about the error number.
2812 * /etc/crypttab entries now understand the new keyfile-size=
2815 * shutdown(8) now can send a (configurable) wall message when
2816 a shutdown is cancelled.
2818 * The mount propagation mode for the root file system will now
2819 default to "shared", which is useful to make containers work
2820 nicely out-of-the-box so that they receive new mounts from
2821 the host. This can be undone locally by running "mount
2822 --make-rprivate /" if needed.
2824 * The prefdm.service file has been removed. Distributions
2825 should maintain this unit downstream if they intend to keep
2826 it around. However, we recommend writing normal unit files
2827 for display managers instead.
2829 * Since systemd is a crucial part of the OS we will now
2830 default to a number of compiler switches that improve
2831 security (hardening) such as read-only relocations, stack
2832 protection, and suchlike.
2834 * The TimeoutSec= setting for services is now split into
2835 TimeoutStartSec= and TimeoutStopSec= to allow configuration
2836 of individual time outs for the start and the stop phase of
2839 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
2840 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
2841 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
2842 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
2843 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
2844 Gundersen, Zbigniew Jędrzejewski-Szmek
2848 * The journal and id128 C APIs are now fully documented as man
2851 * Extra safety checks have been added when transitioning from
2852 the initial RAM disk to the main system to avoid accidental
2855 * /etc/crypttab entries now understand the new keyfile-offset=
2858 * systemctl -t can now be used to filter by unit load state.
2860 * The journal C API gained the new sd_journal_wait() call to
2861 make writing synchronous journal clients easier.
2863 * journalctl gained the new -D switch to show journals from a
2866 * journalctl now displays a special marker between log
2867 messages of two different boots.
2869 * The journal is now explicitly flushed to /var via a service
2870 systemd-journal-flush.service, rather than implicitly simply
2871 by seeing /var/log/journal to be writable.
2873 * journalctl (and the journal C APIs) can now match for much
2874 more complex expressions, with alternatives and
2877 * When transitioning from the initial RAM disk to the main
2878 system we will now kill all processes in a killing spree to
2879 ensure no processes stay around by accident.
2881 * Three new specifiers may be used in unit files: %u, %h, %s
2882 resolve to the user name, user home directory resp. user
2883 shell. This is useful for running systemd user instances.
2885 * We now automatically rotate journal files if their data
2886 object hash table gets a fill level > 75%. We also size the
2887 hash table based on the configured maximum file size. This
2888 together should lower hash collisions drastically and thus
2889 speed things up a bit.
2891 * journalctl gained the new "--header" switch to introspect
2892 header data of journal files.
2894 * A new setting SystemCallFilters= has been added to services
2895 which may be used to apply blacklists or whitelists to
2896 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
2898 * nspawn gained a new --link-journal= switch (and quicker: -j)
2899 to link the container journal with the host. This makes it
2900 very easy to centralize log viewing on the host for all
2901 guests while still keeping the journal files separated.
2903 * Many bugfixes and optimizations
2905 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
2906 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
2907 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
2912 * Several tools now understand kernel command line arguments,
2913 which are only read when run in an initial RAM disk. They
2914 usually follow closely their normal counterparts, but are
2917 * There's a new tool to analyze the readahead files that are
2918 automatically generated at boot. Use:
2920 /usr/lib/systemd/systemd-readahead analyze /.readahead
2922 * We now provide an early debug shell on tty9 if this enabled. Use:
2924 systemctl enable debug-shell.service
2926 * All plymouth related units have been moved into the Plymouth
2927 package. Please make sure to upgrade your Plymouth version
2930 * systemd-tmpfiles now supports getting passed the basename of
2931 a configuration file only, in which case it will look for it
2932 in all appropriate directories automatically.
2934 * udevadm info now takes a /dev or /sys path as argument, and
2935 does the right thing. Example:
2937 udevadm info /dev/sda
2938 udevadm info /sys/class/block/sda
2940 * systemctl now prints a warning if a unit is stopped but a
2941 unit that might trigger it continues to run. Example: a
2942 service is stopped but the socket that activates it is left
2945 * "systemctl status" will now mention if the log output was
2946 shortened due to rotation since a service has been started.
2948 * The journal API now exposes functions to determine the
2949 "cutoff" times due to rotation.
2951 * journald now understands SIGUSR1 and SIGUSR2 for triggering
2952 immediately flushing of runtime logs to /var if possible,
2953 resp. for triggering immediate rotation of the journal
2956 * It is now considered an error if a service is attempted to
2957 be stopped that is not loaded.
2959 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
2961 * systemd-analyze now supports Python 3
2963 * tmpfiles now supports cleaning up directories via aging
2964 where the first level dirs are always kept around but
2965 directories beneath it automatically aged. This is enabled
2966 by prefixing the age field with '~'.
2968 * Seat objects now expose CanGraphical, CanTTY properties
2969 which is required to deal with very fast bootups where the
2970 display manager might be running before the graphics drivers
2971 completed initialization.
2973 * Seat objects now expose a State property.
2975 * We now include RPM macros for service enabling/disabling
2976 based on the preset logic. We recommend RPM based
2977 distributions to make use of these macros if possible. This
2978 makes it simpler to reuse RPM spec files across
2981 * We now make sure that the collected systemd unit name is
2982 always valid when services log to the journal via
2985 * There's a new man page kernel-command-line(7) detailing all
2986 command line options we understand.
2988 * The fstab generator may now be disabled at boot by passing
2989 fstab=0 on the kernel command line.
2991 * A new kernel command line option modules-load= is now understood
2992 to load a specific kernel module statically, early at boot.
2994 * Unit names specified on the systemctl command line are now
2995 automatically escaped as needed. Also, if file system or
2996 device paths are specified they are automatically turned
2997 into the appropriate mount or device unit names. Example:
2999 systemctl status /home
3000 systemctl status /dev/sda
3002 * The SysVConsole= configuration option has been removed from
3003 system.conf parsing.
3005 * The SysV search path is no longer exported on the D-Bus
3008 * The Names= option is been removed from unit file parsing.
3010 * There's a new man page bootup(7) detailing the boot process.
3012 * Every unit and every generator we ship with systemd now
3013 comes with full documentation. The self-explanatory boot is
3016 * A couple of services gained "systemd-" prefixes in their
3017 name if they wrap systemd code, rather than only external
3018 code. Among them fsck@.service which is now
3019 systemd-fsck@.service.
3021 * The HaveWatchdog property has been removed from the D-Bus
3024 * systemd.confirm_spawn= on the kernel command line should now
3027 * There's a new man page crypttab(5) which details all options
3028 we actually understand.
3030 * systemd-nspawn gained a new --capability= switch to pass
3031 additional capabilities to the container.
3033 * timedated will now read known NTP implementation unit names
3034 from /usr/lib/systemd/ntp-units.d/*.list,
3035 systemd-timedated-ntp.target has been removed.
3037 * journalctl gained a new switch "-b" that lists log data of
3038 the current boot only.
3040 * The notify socket is in the abstract namespace again, in
3041 order to support daemons which chroot() at start-up.
3043 * There is a new Storage= configuration option for journald
3044 which allows configuration of where log data should go. This
3045 also provides a way to disable journal logging entirely, so
3046 that data collected is only forwarded to the console, the
3047 kernel log buffer or another syslog implementation.
3049 * Many bugfixes and optimizations
3051 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3052 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3053 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3054 Shawn Landden, Tom Gundersen
3058 * "systemctl help <unit>" now shows the man page if one is
3061 * Several new man pages have been added.
3063 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3064 MaxLevelConsole= can now be specified in
3065 journald.conf. These options allow reducing the amount of
3066 data stored on disk or forwarded by the log level.
3068 * TimerSlackNSec= can now be specified in system.conf for
3069 PID1. This allows system-wide power savings.
3071 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3072 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3077 * logind is now capable of (optionally) handling power and
3078 sleep keys as well as the lid switch.
3080 * journalctl now understands the syntax "journalctl
3081 /usr/bin/avahi-daemon" to get all log output of a specific
3084 * CapabilityBoundingSet= in system.conf now also influences
3085 the capability bound set of usermode helpers of the kernel.
3087 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3088 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3089 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3090 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3094 * Note that we skipped 139 releases here in order to set the
3095 new version to something that is greater than both udev's
3096 and systemd's most recent version number.
3098 * udev: all udev sources are merged into the systemd source tree now.
3099 All future udev development will happen in the systemd tree. It
3100 is still fully supported to use the udev daemon and tools without
3101 systemd running, like in initramfs or other init systems. Building
3102 udev though, will require the *build* of the systemd tree, but
3103 udev can be properly *run* without systemd.
3105 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3106 should be used to create dead device nodes as workarounds for broken
3109 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3110 no longer supported. udev_monitor_new_from_netlink() needs to be
3111 used to subscribe to events.
3113 * udev: when udevd is started by systemd, processes which are left
3114 behind by forking them off of udev rules, are unconditionally cleaned
3115 up and killed now after the event handling has finished. Services or
3116 daemons must be started as systemd services. Services can be
3117 pulled-in by udev to get started, but they can no longer be directly
3118 forked by udev rules.
3120 * udev: the daemon binary is called systemd-udevd now and installed
3121 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3122 to adapt to that, create symlink, or rename the binary after building
3125 * libudev no longer provides these symbols:
3126 udev_monitor_from_socket()
3127 udev_queue_get_failed_list_entry()
3128 udev_get_{dev,sys,run}_path()
3129 The versions number was bumped and symbol versioning introduced.
3131 * systemd-loginctl and systemd-journalctl have been renamed
3132 to loginctl and journalctl to match systemctl.
3134 * The config files: /etc/systemd/systemd-logind.conf and
3135 /etc/systemd/systemd-journald.conf have been renamed to
3136 logind.conf and journald.conf. Package updates should rename
3137 the files to the new names on upgrade.
3139 * For almost all files the license is now LGPL2.1+, changed
3140 from the previous GPL2.0+. Exceptions are some minor stuff
3141 of udev (which will be changed to LGPL2.1 eventually, too),
3142 and the MIT licensed sd-daemon.[ch] library that is suitable
3143 to be used as drop-in files.
3145 * systemd and logind now handle system sleep states, in
3146 particular suspending and hibernating.
3148 * logind now implements a sleep/shutdown/idle inhibiting logic
3149 suitable for a variety of uses. Soonishly Lennart will blog
3150 about this in more detail.
3152 * var-run.mount and var-lock.mount are no longer provided
3153 (which prevously bind mounted these directories to their new
3154 places). Distributions which have not converted these
3155 directories to symlinks should consider stealing these files
3156 from git history and add them downstream.
3158 * We introduced the Documentation= field for units and added
3159 this to all our shipped units. This is useful to make it
3160 easier to explore the boot and the purpose of the various
3163 * All smaller setup units (such as
3164 systemd-vconsole-setup.service) now detect properly if they
3165 are run in a container and are skipped when
3166 appropriate. This guarantees an entirely noise-free boot in
3167 Linux container environments such as systemd-nspawn.
3169 * A framework for implementing offline system updates is now
3170 integrated, for details see:
3171 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
3173 * A new service type Type=idle is available now which helps us
3174 avoiding ugly interleaving of getty output and boot status
3177 * There's now a system-wide CapabilityBoundingSet= option to
3178 globally reduce the set of capabilities for the
3179 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
3180 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
3181 even CAP_NET_ADMIN system-wide for secure systems.
3183 * There are now system-wide DefaultLimitXXX= options to
3184 globally change the defaults of the various resource limits
3185 for all units started by PID 1.
3187 * Harald Hoyer's systemd test suite has been integrated into
3188 systemd which allows easy testing of systemd builds in qemu
3189 and nspawn. (This is really awesome! Ask us for details!)
3191 * The fstab parser is now implemented as generator, not inside
3194 * systemctl will now warn you if .mount units generated from
3195 /etc/fstab are out of date due to changes in fstab that
3196 have not been read by systemd yet.
3198 * systemd is now suitable for usage in initrds. Dracut has
3199 already been updated to make use of this. With this in place
3200 initrds get a slight bit faster but primarily are much
3201 easier to introspect and debug since "systemctl status" in
3202 the host system can be used to introspect initrd services,
3203 and the journal from the initrd is kept around too.
3205 * systemd-delta has been added, a tool to explore differences
3206 between user/admin configuration and vendor defaults.
3208 * PrivateTmp= now affects both /tmp and /var/tmp.
3210 * Boot time status messages are now much prettier and feature
3211 proper english language. Booting up systemd has never been
3214 * Read-ahead pack files now include the inode number of all
3215 files to pre-cache. When the inode changes the pre-caching
3216 is not attempted. This should be nicer to deal with updated
3217 packages which might result in changes of read-ahead
3220 * We now temporaritly lower the kernel's read_ahead_kb variable
3221 when collecting read-ahead data to ensure the kernel's
3222 built-in read-ahead does not add noise to our measurements
3223 of necessary blocks to pre-cache.
3225 * There's now RequiresMountsFor= to add automatic dependencies
3226 for all mounts necessary for a specific file system path.
3228 * MountAuto= and SwapAuto= have been removed from
3229 system.conf. Mounting file systems at boot has to take place
3232 * nspawn now learned a new switch --uuid= to set the machine
3233 ID on the command line.
3235 * nspawn now learned the -b switch to automatically search
3238 * vt102 is now the default TERM for serial TTYs, upgraded from
3241 * systemd-logind now works on VT-less systems.
3243 * The build tree has been reorganized. The individual
3244 components now have directories of their own.
3246 * A new condition type ConditionPathIsReadWrite= is now available.
3248 * nspawn learned the new -C switch to create cgroups for the
3249 container in other hierarchies.
3251 * We now have support for hardware watchdogs, configurable in
3254 * The scheduled shutdown logic now has a public API.
3256 * We now mount /tmp as tmpfs by default, but this can be
3257 masked and /etc/fstab can override it.
3259 * Since udisks does not make use of /media anymore we are not
3260 mounting a tmpfs on it anymore.
3262 * journalctl gained a new --local switch to only interleave
3263 locally generated journal files.
3265 * We can now load the IMA policy at boot automatically.
3267 * The GTK tools have been split off into a systemd-ui.
3269 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3270 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3271 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3272 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3273 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3274 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3275 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3276 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3277 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3282 * This is mostly a bugfix release
3284 * Support optional initialization of the machine ID from the
3285 KVM or container configured UUID.
3287 * Support immediate reboots with "systemctl reboot -ff"
3289 * Show /etc/os-release data in systemd-analyze output
3291 * Many bugfixes for the journal, including endianness fixes and
3292 ensuring that disk space enforcement works
3294 * sd-login.h is C++ comptaible again
3296 * Extend the /etc/os-release format on request of the Debian
3299 * We now refuse non-UTF8 strings used in various configuration
3300 and unit files. This is done to ensure we do not pass invalid
3301 data over D-Bus or expose it elsewhere.
3303 * Register Mimo USB Screens as suitable for automatic seat
3306 * Read SELinux client context from journal clients in a race
3309 * Reorder configuration file lookup order. /etc now always
3310 overrides /run in order to allow the administrator to always
3311 and unconditionally override vendor supplied or
3312 automatically generated data.
3314 * The various user visible bits of the journal now have man
3315 pages. We still lack man pages for the journal API calls
3318 * We now ship all man pages in HTML format again in the
3321 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3322 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3323 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3328 * This is mostly a bugfix release
3330 * systems lacking /etc/os-release are no longer supported.
3332 * Various functionality updates to libsystemd-login.so
3334 * Track class of PAM logins to distuingish greeters from
3337 Contributions from: Kay Sievers, Lennart Poettering, Michael
3342 * This is an important bugfix release for v41.
3344 * Building man pages is now optional which should be useful
3345 for those building systemd from git but unwilling to install
3348 * Watchdog support for supervising services is now usable. In
3349 a future release support for hardware watchdogs
3350 (i.e. /dev/watchdog) will be added building on this.
3352 * Service start rate limiting is now configurable and can be
3353 turned off per service. When a start rate limit is hit a
3354 reboot can automatically be triggered.
3356 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
3358 Contributions from: Benjamin Franzke, Bill Nottingham,
3359 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
3360 Schmidt, Michał Górny, Piotr Drąg
3364 * The systemd binary is installed /usr/lib/systemd/systemd now;
3365 An existing /sbin/init symlink needs to be adapted with the
3368 * The code that loads kernel modules has been ported to invoke
3369 libkmod directly, instead of modprobe. This means we do not
3370 support systems with module-init-tools anymore.
3372 * Watchdog support is now already useful, but still not
3375 * A new kernel command line option systemd.setenv= is
3376 understood to set system wide environment variables
3377 dynamically at boot.
3379 * We now limit the set of capabilities of systemd-journald.
3381 * We now set SIGPIPE to ignore by default, since it only is
3382 useful in shell pipelines, and has little use in general
3383 code. This can be disabled with IgnoreSIPIPE=no in unit
3386 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
3387 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
3392 * This is mostly a bugfix release
3394 * We now expose the reason why a service failed in the
3395 "Result" D-Bus property.
3397 * Rudimentary service watchdog support (will be completed over
3398 the next few releases.)
3400 * When systemd forks off in order execute some service we will
3401 now immediately changes its argv[0] to reflect which process
3402 it will execute. This is useful to minimize the time window
3403 with a generic argv[0], which makes bootcharts more useful
3405 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
3406 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
3407 Mike Kazantsev, Ray Strode
3411 * This is mostly a test release, but incorporates many
3414 * New systemd-cgtop tool to show control groups by their
3417 * Linking against libacl for ACLs is optional again. If
3418 disabled, support tracking device access for active logins
3419 goes becomes unavailable, and so does access to the user
3420 journals by the respective users.
3422 * If a group "adm" exists, journal files are automatically
3423 owned by them, thus allow members of this group full access
3424 to the system journal as well as all user journals.
3426 * The journal now stores the SELinux context of the logging
3427 client for all entries.
3429 * Add C++ inclusion guards to all public headers
3431 * New output mode "cat" in the journal to print only text
3432 messages, without any meta data like date or time.
3434 * Include tiny X server wrapper as a temporary stop-gap to
3435 teach XOrg udev display enumeration. This is used by display
3436 managers such as gdm, and will go away as soon as XOrg
3437 learned native udev hotplugging for display devices.
3439 * Add new systemd-cat tool for executing arbitrary programs
3440 with STDERR/STDOUT connected to the journal. Can also act as
3441 BSD logger replacement, and does so by default.
3443 * Optionally store all locally generated coredumps in the
3444 journal along with meta data.
3446 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
3447 writing short strings to files (for usage for /sys), and for
3448 creating symlinks, character and block device nodes.
3450 * New unit file option ControlGroupPersistent= to make cgroups
3451 persistent, following the mechanisms outlined in
3452 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
3454 * Support multiple local RTCs in a sane way
3456 * No longer monopolize IO when replaying readahead data on
3457 rotating disks, since we might starve non-file-system IO to
3458 death, since fanotify() will not see accesses done by blkid,
3461 * Do not show kernel threads in systemd-cgls anymore, unless
3462 requested with new -k switch.
3464 Contributions from: Dan Horák, Kay Sievers, Lennart
3465 Poettering, Michal Schmidt
3469 * This is mostly a test release, but incorporates many
3472 * The git repository moved to:
3473 git://anongit.freedesktop.org/systemd/systemd
3474 ssh://git.freedesktop.org/git/systemd/systemd
3476 * First release with the journal
3477 http://0pointer.de/blog/projects/the-journal.html
3479 * The journal replaces both systemd-kmsg-syslogd and
3480 systemd-stdout-bridge.
3482 * New sd_pid_get_unit() API call in libsystemd-logind
3484 * Many systemadm clean-ups
3486 * Introduce remote-fs-pre.target which is ordered before all
3487 remote mounts and may be used to start services before all
3490 * Added Mageia support
3492 * Add bash completion for systemd-loginctl
3494 * Actively monitor PID file creation for daemons which exit in
3495 the parent process before having finished writing the PID
3496 file in the daemon process. Daemons which do this need to be
3497 fixed (i.e. PID file creation must have finished before the
3498 parent exits), but we now react a bit more gracefully to them.
3500 * Add colourful boot output, mimicking the well-known output
3501 of existing distributions.
3503 * New option PassCredentials= for socket units, for
3504 compatibility with a recent kernel ABI breakage.
3506 * /etc/rc.local is now hooked in via a generator binary, and
3507 thus will no longer act as synchronization point during
3510 * systemctl list-unit-files now supports --root=.
3512 * systemd-tmpfiles now understands two new commands: z, Z for
3513 relabelling files according to the SELinux database. This is
3514 useful to apply SELinux labels to specific files in /sys,
3517 * Output of SysV services is now forwarded to both the console
3518 and the journal by default, not only just the console.
3520 * New man pages for all APIs from libsystemd-login.
3522 * The build tree got reorganized and a the build system is a
3523 lot more modular allowing embedded setups to specifically
3524 select the components of systemd they are interested in.
3526 * Support for Linux systems lacking the kernel VT subsystem is
3529 * configure's --with-rootdir= got renamed to
3530 --with-rootprefix= to follow the naming used by udev and
3533 * Unless specified otherwise we will now install to /usr instead
3534 of /usr/local by default.
3536 * Processes with '@' in argv[0][0] are now excluded from the
3537 final shut-down killing spree, following the logic explained
3539 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
3541 * All processes remaining in a service cgroup when we enter
3542 the START or START_PRE states are now killed with
3543 SIGKILL. That means it is no longer possible to spawn
3544 background processes from ExecStart= lines (which was never
3545 supported anyway, and bad style).
3547 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
3548 reloading of units together.
3550 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
3551 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
3552 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
3553 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
3554 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek